3 * @file src/Network/Probe.php
5 namespace Friendica\Network;
8 * @file src/Network/Probe.php
9 * @brief Functions for probing URL
13 use Friendica\Core\System;
14 use Friendica\Core\Cache;
15 use Friendica\Core\Config;
16 use Friendica\Database\DBM;
17 use Friendica\Model\Profile;
18 use Friendica\Protocol\Email;
19 use Friendica\Protocol\Feed;
20 use Friendica\Util\Crypto;
21 use Friendica\Util\Network;
22 use Friendica\Util\XML;
28 require_once 'include/dba.php';
31 * @brief This class contain functions for probing URL
36 private static $baseurl;
39 * @brief Rearrange the array so that it always has the same order
41 * @param array $data Unordered data
43 * @return array Ordered data
45 private static function rearrangeData($data)
47 $fields = ["name", "nick", "guid", "url", "addr", "alias",
48 "photo", "community", "keywords", "location", "about",
49 "batch", "notify", "poll", "request", "confirm", "poco",
50 "priority", "network", "pubkey", "baseurl"];
53 foreach ($fields as $field) {
54 if (isset($data[$field])) {
55 $newdata[$field] = $data[$field];
57 $newdata[$field] = "";
61 // We don't use the "priority" field anymore and replace it with a dummy.
62 $newdata["priority"] = 0;
68 * @brief Check if the hostname belongs to the own server
70 * @param string $host The hostname that is to be checked
72 * @return bool Does the testes hostname belongs to the own server?
74 private static function ownHost($host)
76 $own_host = get_app()->get_hostname();
78 $parts = parse_url($host);
80 if (!isset($parts['scheme'])) {
81 $parts = parse_url('http://'.$host);
84 if (!isset($parts['host'])) {
87 return $parts['host'] == $own_host;
91 * @brief Probes for webfinger path via "host-meta"
93 * We have to check if the servers in the future still will offer this.
94 * It seems as if it was dropped from the standard.
96 * @param string $host The host part of an url
98 * @return array with template and type of the webfinger template for JSON or XML
100 private static function hostMeta($host)
102 // Reset the static variable
105 $ssl_url = "https://".$host."/.well-known/host-meta";
106 $url = "http://".$host."/.well-known/host-meta";
108 $xrd_timeout = Config::get('system', 'xrd_timeout', 20);
111 logger("Probing for ".$host, LOGGER_DEBUG);
113 $ret = Network::curl($ssl_url, false, $redirects, ['timeout' => $xrd_timeout, 'accept_content' => 'application/xrd+xml']);
114 if ($ret['success']) {
116 $xrd = XML::parseString($xml, false);
117 $host_url = 'https://'.$host;
120 if (!is_object($xrd)) {
121 $ret = Network::curl($url, false, $redirects, ['timeout' => $xrd_timeout, 'accept_content' => 'application/xrd+xml']);
122 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
123 logger("Probing timeout for ".$url, LOGGER_DEBUG);
127 $xrd = XML::parseString($xml, false);
128 $host_url = 'http://'.$host;
130 if (!is_object($xrd)) {
131 logger("No xrd object found for ".$host, LOGGER_DEBUG);
135 $links = XML::elementToArray($xrd);
136 if (!isset($links["xrd"]["link"])) {
137 logger("No xrd data found for ".$host, LOGGER_DEBUG);
142 // The following webfinger path is defined in RFC 7033 https://tools.ietf.org/html/rfc7033
143 // Problem is that Hubzilla currently doesn't provide all data in the JSON webfinger
144 // compared to the XML webfinger. So this is commented out by now.
145 // $lrdd = array("application/jrd+json" => $host_url.'/.well-known/webfinger?resource={uri}');
147 foreach ($links["xrd"]["link"] as $value => $link) {
148 if (!empty($link["@attributes"])) {
149 $attributes = $link["@attributes"];
150 } elseif ($value == "@attributes") {
156 if (($attributes["rel"] == "lrdd") && !empty($attributes["template"])) {
157 $type = (empty($attributes["type"]) ? '' : $attributes["type"]);
159 $lrdd[$type] = $attributes["template"];
163 self::$baseurl = "http://".$host;
165 logger("Probing successful for ".$host, LOGGER_DEBUG);
171 * @brief Perform Webfinger lookup and return DFRN data
173 * Given an email style address, perform webfinger lookup and
174 * return the resulting DFRN profile URL, or if no DFRN profile URL
175 * is located, returns an OStatus subscription template (prefixed
176 * with the string 'stat:' to identify it as on OStatus template).
177 * If this isn't an email style address just return $webbie.
178 * Return an empty string if email-style addresses but webfinger fails,
179 * or if the resultant personal XRD doesn't contain a supported
180 * subscription/friend-request attribute.
182 * amended 7/9/2011 to return an hcard which could save potentially loading
183 * a lengthy content page to scrape dfrn attributes
185 * @param string $webbie Address that should be probed
186 * @param string $hcard_url Link to the hcard - is returned by reference
188 * @return string profile link
190 public static function webfingerDfrn($webbie, &$hcard_url)
194 $links = self::lrdd($webbie);
195 logger('webfingerDfrn: '.$webbie.':'.print_r($links, true), LOGGER_DATA);
197 foreach ($links as $link) {
198 if ($link['@attributes']['rel'] === NAMESPACE_DFRN) {
199 $profile_link = $link['@attributes']['href'];
201 if (($link['@attributes']['rel'] === NAMESPACE_OSTATUSSUB) && ($profile_link == "")) {
202 $profile_link = 'stat:'.$link['@attributes']['template'];
204 if ($link['@attributes']['rel'] === 'http://microformats.org/profile/hcard') {
205 $hcard_url = $link['@attributes']['href'];
209 return $profile_link;
213 * @brief Check an URI for LRDD data
215 * this is a replacement for the "lrdd" function.
216 * It isn't used in this class and has some redundancies in the code.
217 * When time comes we can check the existing calls for "lrdd" if we can rework them.
219 * @param string $uri Address that should be probed
221 * @return array uri data
223 public static function lrdd($uri)
225 $lrdd = self::hostMeta($uri);
228 if (is_bool($lrdd)) {
233 $parts = @parse_url($uri);
238 $host = $parts["host"];
239 if (!empty($parts["port"])) {
240 $host .= ':'.$parts["port"];
243 $path_parts = explode("/", trim($parts["path"], "/"));
245 $nick = array_pop($path_parts);
248 $lrdd = self::hostMeta($host);
249 $host .= "/".array_shift($path_parts);
250 } while (!$lrdd && (sizeof($path_parts) > 0));
254 logger("No lrdd data found for ".$uri, LOGGER_DEBUG);
258 foreach ($lrdd as $type => $template) {
263 $path = str_replace('{uri}', urlencode($uri), $template);
264 $webfinger = self::webfinger($path, $type);
266 if (!$webfinger && (strstr($uri, "@"))) {
267 $path = str_replace('{uri}', urlencode("acct:".$uri), $template);
268 $webfinger = self::webfinger($path, $type);
271 // Special treatment for Mastodon
272 // Problem is that Mastodon uses an URL format like http://domain.tld/@nick
273 // But the webfinger for this format fails.
274 if (!$webfinger && !empty($nick)) {
275 // Mastodon uses a "@" as prefix for usernames in their url format
276 $nick = ltrim($nick, '@');
278 $addr = $nick."@".$host;
280 $path = str_replace('{uri}', urlencode("acct:".$addr), $template);
281 $webfinger = self::webfinger($path, $type);
285 if (!is_array($webfinger["links"])) {
286 logger("No webfinger links found for ".$uri, LOGGER_DEBUG);
292 foreach ($webfinger["links"] as $link) {
293 $data[] = ["@attributes" => $link];
296 if (is_array($webfinger["aliases"])) {
297 foreach ($webfinger["aliases"] as $alias) {
298 $data[] = ["@attributes" =>
308 * @brief Fetch information (protocol endpoints and user information) about a given uri
310 * @param string $uri Address that should be probed
311 * @param string $network Test for this specific network
312 * @param integer $uid User ID for the probe (only used for mails)
313 * @param boolean $cache Use cached values?
315 * @return array uri data
317 public static function uri($uri, $network = "", $uid = -1, $cache = true)
320 $result = Cache::get("Probe::uri:".$network.":".$uri);
321 if (!is_null($result)) {
330 $data = self::detect($uri, $network, $uid);
332 if (!isset($data["url"])) {
336 if (x($data, "photo")) {
337 $data["baseurl"] = Network::getUrlMatch(normalise_link($data["baseurl"]), normalise_link($data["photo"]));
339 $data["photo"] = System::baseUrl().'/images/person-175.jpg';
342 if (empty($data["name"])) {
343 if (!empty($data["nick"])) {
344 $data["name"] = $data["nick"];
347 if (!x($data, "name")) {
348 $data["name"] = $data["url"];
352 if (empty($data["nick"])) {
353 $data["nick"] = strtolower($data["name"]);
355 if (strpos($data['nick'], ' ')) {
356 $data['nick'] = trim(substr($data['nick'], 0, strpos($data['nick'], ' ')));
360 if (self::$baseurl != "") {
361 $data["baseurl"] = self::$baseurl;
364 if (!isset($data["network"])) {
365 $data["network"] = NETWORK_PHANTOM;
368 $data = self::rearrangeData($data);
370 // Only store into the cache if the value seems to be valid
371 if (!in_array($data['network'], [NETWORK_PHANTOM, NETWORK_MAIL])) {
372 Cache::set("Probe::uri:".$network.":".$uri, $data, CACHE_DAY);
374 /// @todo temporary fix - we need a real contact update function that updates only changing fields
375 /// The biggest problem is the avatar picture that could have a reduced image size.
376 /// It should only be updated if the existing picture isn't existing anymore.
377 /// We only update the contact when it is no probing for a specific network.
378 if (($data['network'] != NETWORK_FEED)
386 $fields = ['name' => $data['name'],
387 'nick' => $data['nick'],
388 'url' => $data['url'],
389 'addr' => $data['addr'],
390 'photo' => $data['photo'],
391 'keywords' => $data['keywords'],
392 'location' => $data['location'],
393 'about' => $data['about'],
394 'notify' => $data['notify'],
395 'network' => $data['network'],
396 'server_url' => $data['baseurl']];
400 foreach ($fields as $key => $val) {
402 unset($fields[$key]);
404 $fieldnames[] = $key;
408 $fields['updated'] = DBM::date();
410 $condition = ['nurl' => normalise_link($data["url"])];
412 $old_fields = dba::selectFirst('gcontact', $fieldnames, $condition);
414 dba::update('gcontact', $fields, $condition, $old_fields);
416 $fields = ['name' => $data['name'],
417 'nick' => $data['nick'],
418 'url' => $data['url'],
419 'addr' => $data['addr'],
420 'alias' => $data['alias'],
421 'keywords' => $data['keywords'],
422 'location' => $data['location'],
423 'about' => $data['about'],
424 'batch' => $data['batch'],
425 'notify' => $data['notify'],
426 'poll' => $data['poll'],
427 'request' => $data['request'],
428 'confirm' => $data['confirm'],
429 'poco' => $data['poco'],
430 'network' => $data['network'],
431 'success_update' => DBM::date()];
435 foreach ($fields as $key => $val) {
437 unset($fields[$key]);
439 $fieldnames[] = $key;
443 $condition = ['nurl' => normalise_link($data["url"]), 'self' => false, 'uid' => 0];
445 $old_fields = dba::selectFirst('contact', $fieldnames, $condition);
447 // When the contact doesn't exist, the value "true" will trigger an insert
452 dba::update('contact', $fields, $condition, $old_fields);
460 * @brief Switch the scheme of an url between http and https
462 * @param string $url URL
464 * @return string switched URL
466 private static function switchScheme($url)
468 $parts = parse_url($url);
470 if (!isset($parts['scheme'])) {
474 if ($parts['scheme'] == 'http') {
475 $url = str_replace('http://', 'https://', $url);
476 } elseif ($parts['scheme'] == 'https') {
477 $url = str_replace('https://', 'http://', $url);
484 * @brief Checks if a profile url should be OStatus but only provides partial information
486 * @param array $webfinger Webfinger data
487 * @param string $lrdd Path template for webfinger request
488 * @param string $type type
490 * @return array fixed webfinger data
492 private static function fixOStatus($webfinger, $lrdd, $type)
494 if (empty($webfinger['links']) || empty($webfinger['subject'])) {
501 foreach ($webfinger['links'] as $link) {
502 if ($link['rel'] == NAMESPACE_OSTATUSSUB) {
505 if ($link['rel'] == 'magic-public-key') {
510 if (!$is_ostatus || $has_key) {
514 $url = self::switchScheme($webfinger['subject']);
515 $path = str_replace('{uri}', urlencode($url), $lrdd);
516 $webfinger2 = self::webfinger($path, $type);
518 // Is the new webfinger detectable as OStatus?
519 if (self::ostatus($webfinger2, true)) {
520 $webfinger = $webfinger2;
527 * @brief Fetch information (protocol endpoints and user information) about a given uri
529 * This function is only called by the "uri" function that adds caching and rearranging of data.
531 * @param string $uri Address that should be probed
532 * @param string $network Test for this specific network
533 * @param integer $uid User ID for the probe (only used for mails)
535 * @return array uri data
537 private static function detect($uri, $network, $uid)
539 $parts = parse_url($uri);
541 if (!empty($parts["scheme"]) && !empty($parts["host"]) && !empty($parts["path"])) {
542 $host = $parts["host"];
543 if (!empty($parts["port"])) {
544 $host .= ':'.$parts["port"];
547 if ($host == 'twitter.com') {
548 return ["network" => NETWORK_TWITTER];
550 $lrdd = self::hostMeta($host);
552 if (is_bool($lrdd)) {
556 $path_parts = explode("/", trim($parts["path"], "/"));
558 while (!$lrdd && (sizeof($path_parts) > 1)) {
559 $host .= "/".array_shift($path_parts);
560 $lrdd = self::hostMeta($host);
563 logger('No XRD data was found for '.$uri, LOGGER_DEBUG);
564 return self::feed($uri);
566 $nick = array_pop($path_parts);
568 // Mastodon uses a "@" as prefix for usernames in their url format
569 $nick = ltrim($nick, '@');
571 $addr = $nick."@".$host;
572 } elseif (strstr($uri, '@')) {
573 // If the URI starts with "mailto:" then jump directly to the mail detection
574 if (strpos($uri, 'mailto:') !== false) {
575 $uri = str_replace('mailto:', '', $uri);
576 return self::mail($uri, $uid);
579 if ($network == NETWORK_MAIL) {
580 return self::mail($uri, $uid);
582 // Remove "acct:" from the URI
583 $uri = str_replace('acct:', '', $uri);
585 $host = substr($uri, strpos($uri, '@') + 1);
586 $nick = substr($uri, 0, strpos($uri, '@'));
588 if (strpos($uri, '@twitter.com')) {
589 return ["network" => NETWORK_TWITTER];
591 $lrdd = self::hostMeta($host);
593 if (is_bool($lrdd)) {
598 logger('No XRD data was found for '.$uri, LOGGER_DEBUG);
599 return self::mail($uri, $uid);
603 logger("Uri ".$uri." was not detectable", LOGGER_DEBUG);
609 /// @todo Do we need the prefix "acct:" or "acct://"?
611 foreach ($lrdd as $type => $template) {
616 // At first try it with the given uri
617 $path = str_replace('{uri}', urlencode($uri), $template);
618 $webfinger = self::webfinger($path, $type);
620 // Fix possible problems with GNU Social probing to wrong scheme
621 $webfinger = self::fixOStatus($webfinger, $template, $type);
623 // We cannot be sure that the detected address was correct, so we don't use the values
624 if ($webfinger && ($uri != $addr)) {
629 // Try webfinger with the address (user@domain.tld)
631 $path = str_replace('{uri}', urlencode($addr), $template);
632 $webfinger = self::webfinger($path, $type);
635 // Mastodon needs to have it with "acct:"
637 $path = str_replace('{uri}', urlencode("acct:".$addr), $template);
638 $webfinger = self::webfinger($path, $type);
643 return self::feed($uri);
648 logger("Probing ".$uri, LOGGER_DEBUG);
650 if (in_array($network, ["", NETWORK_DFRN])) {
651 $result = self::dfrn($webfinger);
653 if ((!$result && ($network == "")) || ($network == NETWORK_DIASPORA)) {
654 $result = self::diaspora($webfinger);
656 if ((!$result && ($network == "")) || ($network == NETWORK_OSTATUS)) {
657 $result = self::ostatus($webfinger);
659 if ((!$result && ($network == "")) || ($network == NETWORK_PUMPIO)) {
660 $result = self::pumpio($webfinger, $addr);
662 if ((!$result && ($network == "")) || ($network == NETWORK_FEED)) {
663 $result = self::feed($uri);
665 // We overwrite the detected nick with our try if the previois routines hadn't detected it.
666 // Additionally it is overwritten when the nickname doesn't make sense (contains spaces).
667 if ((empty($result["nick"]) || (strstr($result["nick"], " "))) && ($nick != "")) {
668 $result["nick"] = $nick;
671 if (empty($result["addr"]) && ($addr != "")) {
672 $result["addr"] = $addr;
676 logger($uri." is ".$result["network"], LOGGER_DEBUG);
678 if (empty($result["baseurl"])) {
679 $pos = strpos($result["url"], $host);
681 $result["baseurl"] = substr($result["url"], 0, $pos).$host;
688 * @brief Perform a webfinger request.
690 * For details see RFC 7033: <https://tools.ietf.org/html/rfc7033>
692 * @param string $url Address that should be probed
693 * @param string $type type
695 * @return array webfinger data
697 private static function webfinger($url, $type)
699 $xrd_timeout = Config::get('system', 'xrd_timeout', 20);
702 $ret = Network::curl($url, false, $redirects, ['timeout' => $xrd_timeout, 'accept_content' => $type]);
703 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
706 $data = $ret['body'];
708 $webfinger = json_decode($data, true);
709 if (is_array($webfinger)) {
710 if (!isset($webfinger["links"])) {
711 logger("No json webfinger links for ".$url, LOGGER_DEBUG);
717 // If it is not JSON, maybe it is XML
718 $xrd = XML::parseString($data, false);
719 if (!is_object($xrd)) {
720 logger("No webfinger data retrievable for ".$url, LOGGER_DEBUG);
724 $xrd_arr = XML::elementToArray($xrd);
725 if (!isset($xrd_arr["xrd"]["link"])) {
726 logger("No XML webfinger links for ".$url, LOGGER_DEBUG);
732 if (!empty($xrd_arr["xrd"]["subject"])) {
733 $webfinger["subject"] = $xrd_arr["xrd"]["subject"];
736 if (!empty($xrd_arr["xrd"]["alias"])) {
737 $webfinger["aliases"] = $xrd_arr["xrd"]["alias"];
740 $webfinger["links"] = [];
742 foreach ($xrd_arr["xrd"]["link"] as $value => $data) {
743 if (!empty($data["@attributes"])) {
744 $attributes = $data["@attributes"];
745 } elseif ($value == "@attributes") {
751 $webfinger["links"][] = $attributes;
757 * @brief Poll the Friendica specific noscrape page.
759 * "noscrape" is a faster alternative to fetch the data from the hcard.
760 * This functionality was originally created for the directory.
762 * @param string $noscrape_url Link to the noscrape page
763 * @param array $data The already fetched data
765 * @return array noscrape data
767 private static function pollNoscrape($noscrape_url, $data)
769 $ret = Network::curl($noscrape_url);
770 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
773 $content = $ret['body'];
775 logger("Empty body for ".$noscrape_url, LOGGER_DEBUG);
779 $json = json_decode($content, true);
780 if (!is_array($json)) {
781 logger("No json data for ".$noscrape_url, LOGGER_DEBUG);
785 if (!empty($json["fn"])) {
786 $data["name"] = $json["fn"];
789 if (!empty($json["addr"])) {
790 $data["addr"] = $json["addr"];
793 if (!empty($json["nick"])) {
794 $data["nick"] = $json["nick"];
797 if (!empty($json["guid"])) {
798 $data["guid"] = $json["guid"];
801 if (!empty($json["comm"])) {
802 $data["community"] = $json["comm"];
805 if (!empty($json["tags"])) {
806 $keywords = implode(" ", $json["tags"]);
807 if ($keywords != "") {
808 $data["keywords"] = $keywords;
812 $location = Profile::formatLocation($json);
814 $data["location"] = $location;
817 if (!empty($json["about"])) {
818 $data["about"] = $json["about"];
821 if (!empty($json["key"])) {
822 $data["pubkey"] = $json["key"];
825 if (!empty($json["photo"])) {
826 $data["photo"] = $json["photo"];
829 if (!empty($json["dfrn-request"])) {
830 $data["request"] = $json["dfrn-request"];
833 if (!empty($json["dfrn-confirm"])) {
834 $data["confirm"] = $json["dfrn-confirm"];
837 if (!empty($json["dfrn-notify"])) {
838 $data["notify"] = $json["dfrn-notify"];
841 if (!empty($json["dfrn-poll"])) {
842 $data["poll"] = $json["dfrn-poll"];
849 * @brief Check for valid DFRN data
851 * @param array $data DFRN data
853 * @return int Number of errors
855 public static function validDfrn($data)
858 if (!isset($data['key'])) {
861 if (!isset($data['dfrn-request'])) {
864 if (!isset($data['dfrn-confirm'])) {
867 if (!isset($data['dfrn-notify'])) {
870 if (!isset($data['dfrn-poll'])) {
877 * @brief Fetch data from a DFRN profile page and via "noscrape"
879 * @param string $profile_link Link to the profile page
881 * @return array profile data
883 public static function profile($profile_link)
887 logger("Check profile ".$profile_link, LOGGER_DEBUG);
889 // Fetch data via noscrape - this is faster
890 $noscrape_url = str_replace(["/hcard/", "/profile/"], "/noscrape/", $profile_link);
891 $data = self::pollNoscrape($noscrape_url, $data);
893 if (!isset($data["notify"])
894 || !isset($data["confirm"])
895 || !isset($data["request"])
896 || !isset($data["poll"])
897 || !isset($data["poco"])
898 || !isset($data["name"])
899 || !isset($data["photo"])
901 $data = self::pollHcard($profile_link, $data, true);
905 $prof_data["addr"] = $data["addr"];
906 $prof_data["nick"] = $data["nick"];
907 $prof_data["dfrn-request"] = $data["request"];
908 $prof_data["dfrn-confirm"] = $data["confirm"];
909 $prof_data["dfrn-notify"] = $data["notify"];
910 $prof_data["dfrn-poll"] = $data["poll"];
911 $prof_data["dfrn-poco"] = $data["poco"];
912 $prof_data["photo"] = $data["photo"];
913 $prof_data["fn"] = $data["name"];
914 $prof_data["key"] = $data["pubkey"];
916 logger("Result for profile ".$profile_link.": ".print_r($prof_data, true), LOGGER_DEBUG);
922 * @brief Check for DFRN contact
924 * @param array $webfinger Webfinger data
926 * @return array DFRN data
928 private static function dfrn($webfinger)
932 foreach ($webfinger["links"] as $link) {
933 if (($link["rel"] == NAMESPACE_DFRN) && ($link["href"] != "")) {
934 $data["network"] = NETWORK_DFRN;
935 } elseif (($link["rel"] == NAMESPACE_FEED) && ($link["href"] != "")) {
936 $data["poll"] = $link["href"];
937 } elseif (($link["rel"] == "http://webfinger.net/rel/profile-page") && ($link["type"] == "text/html") && ($link["href"] != "")) {
938 $data["url"] = $link["href"];
939 } elseif (($link["rel"] == "http://microformats.org/profile/hcard") && ($link["href"] != "")) {
940 $hcard_url = $link["href"];
941 } elseif (($link["rel"] == NAMESPACE_POCO) && ($link["href"] != "")) {
942 $data["poco"] = $link["href"];
943 } elseif (($link["rel"] == "http://webfinger.net/rel/avatar") && ($link["href"] != "")) {
944 $data["photo"] = $link["href"];
945 } elseif (($link["rel"] == "http://joindiaspora.com/seed_location") && ($link["href"] != "")) {
946 $data["baseurl"] = trim($link["href"], '/');
947 } elseif (($link["rel"] == "http://joindiaspora.com/guid") && ($link["href"] != "")) {
948 $data["guid"] = $link["href"];
949 } elseif (($link["rel"] == "diaspora-public-key") && ($link["href"] != "")) {
950 $data["pubkey"] = base64_decode($link["href"]);
952 //if (strstr($data["pubkey"], 'RSA ') || ($link["type"] == "RSA"))
953 if (strstr($data["pubkey"], 'RSA ')) {
954 $data["pubkey"] = Crypto::rsaToPem($data["pubkey"]);
959 if (is_array($webfinger["aliases"])) {
960 foreach ($webfinger["aliases"] as $alias) {
961 if (normalise_link($alias) != normalise_link($data["url"]) && ! strstr($alias, "@")) {
962 $data["alias"] = $alias;
963 } elseif (substr($alias, 0, 5) == 'acct:') {
964 $data["addr"] = substr($alias, 5);
969 if (substr($webfinger["subject"], 0, 5) == "acct:") {
970 $data["addr"] = substr($webfinger["subject"], 5);
973 if (!isset($data["network"]) || ($hcard_url == "")) {
977 // Fetch data via noscrape - this is faster
978 $noscrape_url = str_replace("/hcard/", "/noscrape/", $hcard_url);
979 $data = self::pollNoscrape($noscrape_url, $data);
981 if (isset($data["notify"])
982 && isset($data["confirm"])
983 && isset($data["request"])
984 && isset($data["poll"])
985 && isset($data["name"])
986 && isset($data["photo"])
991 $data = self::pollHcard($hcard_url, $data, true);
997 * @brief Poll the hcard page (Diaspora and Friendica specific)
999 * @param string $hcard_url Link to the hcard page
1000 * @param array $data The already fetched data
1001 * @param boolean $dfrn Poll DFRN specific data
1003 * @return array hcard data
1005 private static function pollHcard($hcard_url, $data, $dfrn = false)
1007 $ret = Network::curl($hcard_url);
1008 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
1011 $content = $ret['body'];
1016 $doc = new DOMDocument();
1017 if (!@$doc->loadHTML($content)) {
1021 $xpath = new DomXPath($doc);
1023 $vcards = $xpath->query("//div[contains(concat(' ', @class, ' '), ' vcard ')]");
1024 if (!is_object($vcards)) {
1028 if ($vcards->length > 0) {
1029 $vcard = $vcards->item(0);
1031 // We have to discard the guid from the hcard in favour of the guid from lrdd
1032 // Reason: Hubzilla doesn't use the value "uid" in the hcard like Diaspora does.
1033 $search = $xpath->query("//*[contains(concat(' ', @class, ' '), ' uid ')]", $vcard); // */
1034 if (($search->length > 0) && ($data["guid"] == "")) {
1035 $data["guid"] = $search->item(0)->nodeValue;
1038 $search = $xpath->query("//*[contains(concat(' ', @class, ' '), ' nickname ')]", $vcard); // */
1039 if ($search->length > 0) {
1040 $data["nick"] = $search->item(0)->nodeValue;
1043 $search = $xpath->query("//*[contains(concat(' ', @class, ' '), ' fn ')]", $vcard); // */
1044 if ($search->length > 0) {
1045 $data["name"] = $search->item(0)->nodeValue;
1048 $search = $xpath->query("//*[contains(concat(' ', @class, ' '), ' searchable ')]", $vcard); // */
1049 if ($search->length > 0) {
1050 $data["searchable"] = $search->item(0)->nodeValue;
1053 $search = $xpath->query("//*[contains(concat(' ', @class, ' '), ' key ')]", $vcard); // */
1054 if ($search->length > 0) {
1055 $data["pubkey"] = $search->item(0)->nodeValue;
1056 if (strstr($data["pubkey"], 'RSA ')) {
1057 $data["pubkey"] = Crypto::rsaToPem($data["pubkey"]);
1061 $search = $xpath->query("//*[@id='pod_location']", $vcard); // */
1062 if ($search->length > 0) {
1063 $data["baseurl"] = trim($search->item(0)->nodeValue, "/");
1068 $photos = $xpath->query("//*[contains(concat(' ', @class, ' '), ' photo ') or contains(concat(' ', @class, ' '), ' avatar ')]", $vcard); // */
1069 foreach ($photos as $photo) {
1071 foreach ($photo->attributes as $attribute) {
1072 $attr[$attribute->name] = trim($attribute->value);
1075 if (isset($attr["src"]) && isset($attr["width"])) {
1076 $avatar[$attr["width"]] = $attr["src"];
1079 // We don't have a width. So we just take everything that we got.
1080 // This is a Hubzilla workaround which doesn't send a width.
1081 if ((sizeof($avatar) == 0) && !empty($attr["src"])) {
1082 $avatar[] = $attr["src"];
1086 if (sizeof($avatar)) {
1088 $data["photo"] = self::fixAvatar(array_pop($avatar), $data["baseurl"]);
1092 // Poll DFRN specific data
1093 $search = $xpath->query("//link[contains(concat(' ', @rel), ' dfrn-')]");
1094 if ($search->length > 0) {
1095 foreach ($search as $link) {
1096 //$data["request"] = $search->item(0)->nodeValue;
1098 foreach ($link->attributes as $attribute) {
1099 $attr[$attribute->name] = trim($attribute->value);
1102 $data[substr($attr["rel"], 5)] = $attr["href"];
1106 // Older Friendica versions had used the "uid" field differently than newer versions
1107 if ($data["nick"] == $data["guid"]) {
1108 unset($data["guid"]);
1117 * @brief Check for Diaspora contact
1119 * @param array $webfinger Webfinger data
1121 * @return array Diaspora data
1123 private static function diaspora($webfinger)
1127 foreach ($webfinger["links"] as $link) {
1128 if (($link["rel"] == "http://microformats.org/profile/hcard") && ($link["href"] != "")) {
1129 $hcard_url = $link["href"];
1130 } elseif (($link["rel"] == "http://joindiaspora.com/seed_location") && ($link["href"] != "")) {
1131 $data["baseurl"] = trim($link["href"], '/');
1132 } elseif (($link["rel"] == "http://joindiaspora.com/guid") && ($link["href"] != "")) {
1133 $data["guid"] = $link["href"];
1134 } elseif (($link["rel"] == "http://webfinger.net/rel/profile-page") && ($link["type"] == "text/html") && ($link["href"] != "")) {
1135 $data["url"] = $link["href"];
1136 } elseif (($link["rel"] == NAMESPACE_FEED) && ($link["href"] != "")) {
1137 $data["poll"] = $link["href"];
1138 } elseif (($link["rel"] == NAMESPACE_POCO) && ($link["href"] != "")) {
1139 $data["poco"] = $link["href"];
1140 } elseif (($link["rel"] == "salmon") && ($link["href"] != "")) {
1141 $data["notify"] = $link["href"];
1142 } elseif (($link["rel"] == "diaspora-public-key") && ($link["href"] != "")) {
1143 $data["pubkey"] = base64_decode($link["href"]);
1145 //if (strstr($data["pubkey"], 'RSA ') || ($link["type"] == "RSA"))
1146 if (strstr($data["pubkey"], 'RSA ')) {
1147 $data["pubkey"] = Crypto::rsaToPem($data["pubkey"]);
1152 if (!isset($data["url"]) || ($hcard_url == "")) {
1156 if (is_array($webfinger["aliases"])) {
1157 foreach ($webfinger["aliases"] as $alias) {
1158 if (normalise_link($alias) != normalise_link($data["url"]) && ! strstr($alias, "@")) {
1159 $data["alias"] = $alias;
1160 } elseif (substr($alias, 0, 5) == 'acct:') {
1161 $data["addr"] = substr($alias, 5);
1166 if (!empty($webfinger["subject"]) && (substr($webfinger["subject"], 0, 5) == 'acct:')) {
1167 $data["addr"] = substr($webfinger["subject"], 5);
1170 // Fetch further information from the hcard
1171 $data = self::pollHcard($hcard_url, $data);
1177 if (isset($data["url"])
1178 && isset($data["guid"])
1179 && isset($data["baseurl"])
1180 && isset($data["pubkey"])
1181 && ($hcard_url != "")
1183 $data["network"] = NETWORK_DIASPORA;
1185 // The Diaspora handle must always be lowercase
1186 $data["addr"] = strtolower($data["addr"]);
1188 // We have to overwrite the detected value for "notify" since Hubzilla doesn't send it
1189 $data["notify"] = $data["baseurl"] . "/receive/users/" . $data["guid"];
1190 $data["batch"] = $data["baseurl"] . "/receive/public";
1199 * @brief Check for OStatus contact
1201 * @param array $webfinger Webfinger data
1202 * @param bool $short Short detection mode
1204 * @return array|bool OStatus data or "false" on error or "true" on short mode
1206 private static function ostatus($webfinger, $short = false)
1210 if (is_array($webfinger["aliases"])) {
1211 foreach ($webfinger["aliases"] as $alias) {
1212 if (strstr($alias, "@") && !strstr(normalise_link($alias), "http://")) {
1213 $data["addr"] = str_replace('acct:', '', $alias);
1218 if (is_string($webfinger["subject"]) && strstr($webfinger["subject"], "@")
1219 && !strstr(normalise_link($webfinger["subject"]), "http://")
1221 $data["addr"] = str_replace('acct:', '', $webfinger["subject"]);
1225 if (is_array($webfinger["links"])) {
1226 foreach ($webfinger["links"] as $link) {
1227 if (($link["rel"] == "http://webfinger.net/rel/profile-page")
1228 && ($link["type"] == "text/html")
1229 && ($link["href"] != "")
1231 $data["url"] = $link["href"];
1232 } elseif (($link["rel"] == "salmon") && ($link["href"] != "")) {
1233 $data["notify"] = $link["href"];
1234 } elseif (($link["rel"] == NAMESPACE_FEED) && ($link["href"] != "")) {
1235 $data["poll"] = $link["href"];
1236 } elseif (($link["rel"] == "magic-public-key") && ($link["href"] != "")) {
1237 $pubkey = $link["href"];
1239 if (substr($pubkey, 0, 5) === 'data:') {
1240 if (strstr($pubkey, ',')) {
1241 $pubkey = substr($pubkey, strpos($pubkey, ',') + 1);
1243 $pubkey = substr($pubkey, 5);
1245 } elseif (normalise_link($pubkey) == 'http://') {
1246 $ret = Network::curl($pubkey);
1247 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
1250 $pubkey = $ret['body'];
1253 $key = explode(".", $pubkey);
1255 if (sizeof($key) >= 3) {
1256 $m = base64url_decode($key[1]);
1257 $e = base64url_decode($key[2]);
1258 $data["pubkey"] = Crypto::meToPem($m, $e);
1264 if (isset($data["notify"]) && isset($data["pubkey"])
1265 && isset($data["poll"])
1266 && isset($data["url"])
1268 $data["network"] = NETWORK_OSTATUS;
1277 // Fetch all additional data from the feed
1278 $ret = Network::curl($data["poll"]);
1279 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
1282 $feed = $ret['body'];
1283 $feed_data = Feed::import($feed, $dummy1, $dummy2, $dummy3, true);
1288 if ($feed_data["header"]["author-name"] != "") {
1289 $data["name"] = $feed_data["header"]["author-name"];
1291 if ($feed_data["header"]["author-nick"] != "") {
1292 $data["nick"] = $feed_data["header"]["author-nick"];
1294 if ($feed_data["header"]["author-avatar"] != "") {
1295 $data["photo"] = self::fixAvatar($feed_data["header"]["author-avatar"], $data["url"]);
1297 if ($feed_data["header"]["author-id"] != "") {
1298 $data["alias"] = $feed_data["header"]["author-id"];
1300 if ($feed_data["header"]["author-location"] != "") {
1301 $data["location"] = $feed_data["header"]["author-location"];
1303 if ($feed_data["header"]["author-about"] != "") {
1304 $data["about"] = $feed_data["header"]["author-about"];
1306 // OStatus has serious issues when the the url doesn't fit (ssl vs. non ssl)
1307 // So we take the value that we just fetched, although the other one worked as well
1308 if ($feed_data["header"]["author-link"] != "") {
1309 $data["url"] = $feed_data["header"]["author-link"];
1312 if (($data['poll'] == $data['url']) && ($data["alias"] != '')) {
1313 $data['url'] = $data["alias"];
1314 $data["alias"] = '';
1317 /// @todo Fetch location and "about" from the feed as well
1322 * @brief Fetch data from a pump.io profile page
1324 * @param string $profile_link Link to the profile page
1326 * @return array profile data
1328 private static function pumpioProfileData($profile_link)
1330 $doc = new DOMDocument();
1331 if (!@$doc->loadHTMLFile($profile_link)) {
1335 $xpath = new DomXPath($doc);
1339 $data["name"] = $xpath->query("//span[contains(@class, 'p-name')]")->item(0)->nodeValue;
1341 if ($data["name"] == '') {
1342 // This is ugly - but pump.io doesn't seem to know a better way for it
1343 $data["name"] = trim($xpath->query("//h1[@class='media-header']")->item(0)->nodeValue);
1344 $pos = strpos($data["name"], chr(10));
1346 $data["name"] = trim(substr($data["name"], 0, $pos));
1350 $data["location"] = $xpath->query("//p[contains(@class, 'p-locality')]")->item(0)->nodeValue;
1352 if ($data["location"] == '') {
1353 $data["location"] = $xpath->query("//p[contains(@class, 'location')]")->item(0)->nodeValue;
1356 $data["about"] = $xpath->query("//p[contains(@class, 'p-note')]")->item(0)->nodeValue;
1358 if ($data["about"] == '') {
1359 $data["about"] = $xpath->query("//p[contains(@class, 'summary')]")->item(0)->nodeValue;
1362 $avatar = $xpath->query("//img[contains(@class, 'u-photo')]")->item(0);
1364 $avatar = $xpath->query("//img[@class='img-rounded media-object']")->item(0);
1367 foreach ($avatar->attributes as $attribute) {
1368 if ($attribute->name == "src") {
1369 $data["photo"] = trim($attribute->value);
1378 * @brief Check for pump.io contact
1380 * @param array $webfinger Webfinger data
1382 * @return array pump.io data
1384 private static function pumpio($webfinger, $addr)
1387 foreach ($webfinger["links"] as $link) {
1388 if (($link["rel"] == "http://webfinger.net/rel/profile-page")
1389 && ($link["type"] == "text/html")
1390 && ($link["href"] != "")
1392 $data["url"] = $link["href"];
1393 } elseif (($link["rel"] == "activity-inbox") && ($link["href"] != "")) {
1394 $data["notify"] = $link["href"];
1395 } elseif (($link["rel"] == "activity-outbox") && ($link["href"] != "")) {
1396 $data["poll"] = $link["href"];
1397 } elseif (($link["rel"] == "dialback") && ($link["href"] != "")) {
1398 $data["dialback"] = $link["href"];
1401 if (isset($data["poll"]) && isset($data["notify"])
1402 && isset($data["dialback"])
1403 && isset($data["url"])
1405 // by now we use these fields only for the network type detection
1406 // So we unset all data that isn't used at the moment
1407 unset($data["dialback"]);
1409 $data["network"] = NETWORK_PUMPIO;
1414 $profile_data = self::pumpioProfileData($data["url"]);
1416 if (!$profile_data) {
1420 $data = array_merge($data, $profile_data);
1422 if (($addr != '') && ($data['name'] != '')) {
1423 $name = trim(str_replace($addr, '', $data['name']));
1425 $data['name'] = $name;
1433 * @brief Check page for feed link
1435 * @param string $url Page link
1437 * @return string feed link
1439 private static function getFeedLink($url)
1441 $doc = new DOMDocument();
1443 if (!@$doc->loadHTMLFile($url)) {
1447 $xpath = new DomXPath($doc);
1449 //$feeds = $xpath->query("/html/head/link[@type='application/rss+xml']");
1450 $feeds = $xpath->query("/html/head/link[@type='application/rss+xml' and @rel='alternate']");
1451 if (!is_object($feeds)) {
1455 if ($feeds->length == 0) {
1461 foreach ($feeds as $feed) {
1463 foreach ($feed->attributes as $attribute) {
1464 $attr[$attribute->name] = trim($attribute->value);
1467 if ($feed_url == "") {
1468 $feed_url = $attr["href"];
1476 * @brief Check for feed contact
1478 * @param string $url Profile link
1479 * @param boolean $probe Do a probe if the page contains a feed link
1481 * @return array feed data
1483 private static function feed($url, $probe = true)
1485 $ret = Network::curl($url);
1486 if ($ret['errno'] == CURLE_OPERATION_TIMEDOUT) {
1489 $feed = $ret['body'];
1490 $feed_data = Feed::import($feed, $dummy1, $dummy2, $dummy3, true);
1497 $feed_url = self::getFeedLink($url);
1503 return self::feed($feed_url, false);
1506 if ($feed_data["header"]["author-name"] != "") {
1507 $data["name"] = $feed_data["header"]["author-name"];
1510 if ($feed_data["header"]["author-nick"] != "") {
1511 $data["nick"] = $feed_data["header"]["author-nick"];
1514 if ($feed_data["header"]["author-avatar"] != "") {
1515 $data["photo"] = $feed_data["header"]["author-avatar"];
1518 if ($feed_data["header"]["author-id"] != "") {
1519 $data["alias"] = $feed_data["header"]["author-id"];
1522 $data["url"] = $url;
1523 $data["poll"] = $url;
1525 if ($feed_data["header"]["author-link"] != "") {
1526 $data["baseurl"] = $feed_data["header"]["author-link"];
1528 $data["baseurl"] = $data["url"];
1531 $data["network"] = NETWORK_FEED;
1537 * @brief Check for mail contact
1539 * @param string $uri Profile link
1540 * @param integer $uid User ID
1542 * @return array mail data
1544 private static function mail($uri, $uid)
1546 if (!Network::isEmailDomainValid($uri)) {
1554 $x = q("SELECT `prvkey` FROM `user` WHERE `uid` = %d LIMIT 1", intval($uid));
1556 $r = q("SELECT * FROM `mailacct` WHERE `uid` = %d AND `server` != '' LIMIT 1", intval($uid));
1558 if (DBM::is_result($x) && DBM::is_result($r)) {
1559 $mailbox = Email::constructMailboxName($r[0]);
1561 openssl_private_decrypt(hex2bin($r[0]['pass']), $password, $x[0]['prvkey']);
1562 $mbox = Email::connect($mailbox, $r[0]['user'], $password);
1568 $msgs = Email::poll($mbox, $uri);
1569 logger('searching '.$uri.', '.count($msgs).' messages found.', LOGGER_DEBUG);
1571 if (!count($msgs)) {
1575 $phost = substr($uri, strpos($uri, '@') + 1);
1578 $data["addr"] = $uri;
1579 $data["network"] = NETWORK_MAIL;
1580 $data["name"] = substr($uri, 0, strpos($uri, '@'));
1581 $data["nick"] = $data["name"];
1582 $data["photo"] = Network::lookupAvatarByEmail($uri);
1583 $data["url"] = 'mailto:'.$uri;
1584 $data["notify"] = 'smtp '.random_string();
1585 $data["poll"] = 'email '.random_string();
1587 $x = Email::messageMeta($mbox, $msgs[0]);
1588 if (stristr($x[0]->from, $uri)) {
1589 $adr = imap_rfc822_parse_adrlist($x[0]->from, '');
1590 } elseif (stristr($x[0]->to, $uri)) {
1591 $adr = imap_rfc822_parse_adrlist($x[0]->to, '');
1594 foreach ($adr as $feadr) {
1595 if ((strcasecmp($feadr->mailbox, $data["name"]) == 0)
1596 &&(strcasecmp($feadr->host, $phost) == 0)
1597 && (strlen($feadr->personal))
1599 $personal = imap_mime_header_decode($feadr->personal);
1601 foreach ($personal as $perspart) {
1602 if ($perspart->charset != "default") {
1603 $data["name"] .= iconv($perspart->charset, 'UTF-8//IGNORE', $perspart->text);
1605 $data["name"] .= $perspart->text;
1609 $data["name"] = notags($data["name"]);
1613 if (!empty($mbox)) {
1621 * @brief Mix two paths together to possibly fix missing parts
1623 * @param string $avatar Path to the avatar
1624 * @param string $base Another path that is hopefully complete
1626 * @return string fixed avatar path
1628 public static function fixAvatar($avatar, $base)
1630 $base_parts = parse_url($base);
1632 // Remove all parts that could create a problem
1633 unset($base_parts['path']);
1634 unset($base_parts['query']);
1635 unset($base_parts['fragment']);
1637 $avatar_parts = parse_url($avatar);
1640 $parts = array_merge($base_parts, $avatar_parts);
1642 // And put them together again
1643 $scheme = isset($parts['scheme']) ? $parts['scheme'] . '://' : '';
1644 $host = isset($parts['host']) ? $parts['host'] : '';
1645 $port = isset($parts['port']) ? ':' . $parts['port'] : '';
1646 $path = isset($parts['path']) ? $parts['path'] : '';
1647 $query = isset($parts['query']) ? '?' . $parts['query'] : '';
1648 $fragment = isset($parts['fragment']) ? '#' . $parts['fragment'] : '';
1650 $fixed = $scheme.$host.$port.$path.$query.$fragment;
1652 logger('Base: '.$base.' - Avatar: '.$avatar.' - Fixed: '.$fixed, LOGGER_DATA);