src/Model/GServer.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2020, Friendica
   4  *
   5  * @license GNU AGPL version 3 or any later version
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as
   9  * published by the Free Software Foundation, either version 3 of the
  10  * License, or (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  19  *
  20  */
  21
  22 namespace Friendica\Model;
  23
  24 use DOMDocument;
  25 use DOMXPath;
  26 use Friendica\Core\Logger;
  27 use Friendica\Core\Protocol;
  28 use Friendica\Core\System;
  29 use Friendica\Core\Worker;
  30 use Friendica\Database\Database;
  31 use Friendica\Database\DBA;
  32 use Friendica\DI;
  33 use Friendica\Module\Register;
  34 use Friendica\Network\CurlResult;
  35 use Friendica\Protocol\Relay;
  36 use Friendica\Util\DateTimeFormat;
  37 use Friendica\Util\Network;
  38 use Friendica\Util\Strings;
  39 use Friendica\Util\XML;
  40
  41 /**
  42  * This class handles GServer related functions
  43  */
  44 class GServer
  45 {
  46         // Directory types
  47         const DT_NONE = 0;
  48         const DT_POCO = 1;
  49         const DT_MASTODON = 2;
  50
  51         // Methods to detect server types
  52
  53         // Non endpoint specific methods
  54         const DETECT_MANUAL = 0;
  55         const DETECT_HEADER = 1;
  56         const DETECT_BODY = 2;
  57
  58         // Implementation specific endpoints
  59         const DETECT_FRIENDIKA = 10;
  60         const DETECT_FRIENDICA = 11;
  61         const DETECT_STATUSNET = 12;
  62         const DETECT_GNUSOCIAL = 13;
  63         const DETECT_CONFIG_JSON = 14; // Statusnet, GNU Social, Older Hubzilla/Redmatrix
  64         const DETECT_SITEINFO_JSON = 15; // Newer Hubzilla
  65         const DETECT_MASTODON_API = 16;
  66         const DETECT_STATUS_PHP = 17; // Nextcloud
  67
  68         // Standardized endpoints
  69         const DETECT_STATISTICS_JSON = 100;
  70         const DETECT_NODEINFO_1 = 101;
  71         const DETECT_NODEINFO_2 = 102;
  72
  73         /**
  74          * Get the ID for the given server URL
  75          *
  76          * @param string $url
  77          * @param boolean $no_check Don't check if the server hadn't been found
  78          * @return int gserver id
  79          */
  80         public static function getID(string $url, bool $no_check = false)
  81         {
  82                 if (empty($url)) {
  83                         return null;
  84                 }
  85
  86                 $url = self::cleanURL($url);
  87
  88                 $gserver = DBA::selectFirst('gserver', ['id'], ['nurl' => Strings::normaliseLink($url)]);
  89                 if (DBA::isResult($gserver)) {
  90                         Logger::info('Got ID for URL', ['id' => $gserver['id'], 'url' => $url, 'callstack' => System::callstack(20)]);
  91                         return $gserver['id'];
  92                 }
  93
  94                 if ($no_check || !self::check($url)) {
  95                         return null;
  96                 }
  97
  98                 return self::getID($url, true);
  99         }
 100
 101         /**
 102          * Checks if the given server is reachable
 103          *
 104          * @param string  $profile URL of the given profile
 105          * @param string  $server  URL of the given server (If empty, taken from profile)
 106          * @param string  $network Network value that is used, when detection failed
 107          * @param boolean $force   Force an update.
 108          *
 109          * @return boolean 'true' if server seems vital
 110          */
 111         public static function reachable(string $profile, string $server = '', string $network = '', bool $force = false)
 112         {
 113                 if ($server == '') {
 114                         $contact = Contact::getByURL($profile, null, ['baseurl']);
 115                         if (!empty($contact['baseurl'])) {
 116                                 $server = $contact['baseurl'];
 117                         }
 118                 }
 119
 120                 if ($server == '') {
 121                         return true;
 122                 }
 123
 124                 return self::check($server, $network, $force);
 125         }
 126
 127         public static function getNextUpdateDate(bool $success, string $created = '', string $last_contact = '')
 128         {
 129                 // On successful contact process check again next week
 130                 if ($success) {
 131                         return DateTimeFormat::utc('now +7 day');
 132                 }
 133
 134                 $now = strtotime(DateTimeFormat::utcNow());
 135
 136                 if ($created > $last_contact) {
 137                         $contact_time = strtotime($created);
 138                 } else {
 139                         $contact_time = strtotime($last_contact);
 140                 }
 141
 142                 // If the last contact was less than 6 hours before then try again in 6 hours
 143                 if (($now - $contact_time) < (60 * 60 * 6)) {
 144                         return DateTimeFormat::utc('now +6 hour');
 145                 }
 146
 147                 // If the last contact was less than 12 hours before then try again in 12 hours
 148                 if (($now - $contact_time) < (60 * 60 * 12)) {
 149                         return DateTimeFormat::utc('now +12 hour');
 150                 }
 151
 152                 // If the last contact was less than 24 hours before then try tomorrow again
 153                 if (($now - $contact_time) < (60 * 60 * 24)) {
 154                         return DateTimeFormat::utc('now +1 day');
 155                 }
 156
 157                 // If the last contact was less than a week before then try again in a week
 158                 if (($now - $contact_time) < (60 * 60 * 24 * 7)) {
 159                         return DateTimeFormat::utc('now +1 week');
 160                 }
 161
 162                 // If the last contact was less than two weeks before then try again in two week
 163                 if (($now - $contact_time) < (60 * 60 * 24 * 14)) {
 164                         return DateTimeFormat::utc('now +2 week');
 165                 }
 166
 167                 // If the last contact was less than a month before then try again in a month
 168                 if (($now - $contact_time) < (60 * 60 * 24 * 30)) {
 169                         return DateTimeFormat::utc('now +1 month');
 170                 }
 171
 172                 // The system hadn't been successul contacted for more than a month, so try again in three months
 173                 return DateTimeFormat::utc('now +3 month');
 174         }
 175
 176         /**
 177          * Decides if a server needs to be updated, based upon several date fields
 178          *
 179          * @param date $created      Creation date of that server entry
 180          * @param date $updated      When had the server entry be updated
 181          * @param date $last_failure Last failure when contacting that server
 182          * @param date $last_contact Last time the server had been contacted
 183          *
 184          * @return boolean Does the server record needs an update?
 185          */
 186         public static function updateNeeded($created, $updated, $last_failure, $last_contact)
 187         {
 188                 $now = strtotime(DateTimeFormat::utcNow());
 189
 190                 if ($updated > $last_contact) {
 191                         $contact_time = strtotime($updated);
 192                 } else {
 193                         $contact_time = strtotime($last_contact);
 194                 }
 195
 196                 $failure_time = strtotime($last_failure);
 197                 $created_time = strtotime($created);
 198
 199                 // If there is no "created" time then use the current time
 200                 if ($created_time <= 0) {
 201                         $created_time = $now;
 202                 }
 203
 204                 // If the last contact was less than 24 hours then don't update
 205                 if (($now - $contact_time) < (60 * 60 * 24)) {
 206                         return false;
 207                 }
 208
 209                 // If the last failure was less than 24 hours then don't update
 210                 if (($now - $failure_time) < (60 * 60 * 24)) {
 211                         return false;
 212                 }
 213
 214                 // If the last contact was less than a week ago and the last failure is older than a week then don't update
 215                 //if ((($now - $contact_time) < (60 * 60 * 24 * 7)) && ($contact_time > $failure_time))
 216                 //      return false;
 217
 218                 // If the last contact time was more than a week ago and the contact was created more than a week ago, then only try once a week
 219                 if ((($now - $contact_time) > (60 * 60 * 24 * 7)) && (($now - $created_time) > (60 * 60 * 24 * 7)) && (($now - $failure_time) < (60 * 60 * 24 * 7))) {
 220                         return false;
 221                 }
 222
 223                 // If the last contact time was more than a month ago and the contact was created more than a month ago, then only try once a month
 224                 if ((($now - $contact_time) > (60 * 60 * 24 * 30)) && (($now - $created_time) > (60 * 60 * 24 * 30)) && (($now - $failure_time) < (60 * 60 * 24 * 30))) {
 225                         return false;
 226                 }
 227
 228                 return true;
 229         }
 230
 231         /**
 232          * Checks the state of the given server.
 233          *
 234          * @param string  $server_url    URL of the given server
 235          * @param string  $network       Network value that is used, when detection failed
 236          * @param boolean $force         Force an update.
 237          * @param boolean $only_nodeinfo Only use nodeinfo for server detection
 238          *
 239          * @return boolean 'true' if server seems vital
 240          */
 241         public static function check(string $server_url, string $network = '', bool $force = false, bool $only_nodeinfo = false)
 242         {
 243                 $server_url = self::cleanURL($server_url);
 244
 245                 if ($server_url == '') {
 246                         return false;
 247                 }
 248
 249                 $gserver = DBA::selectFirst('gserver', [], ['nurl' => Strings::normaliseLink($server_url)]);
 250                 if (DBA::isResult($gserver)) {
 251                         if ($gserver['created'] <= DBA::NULL_DATETIME) {
 252                                 $fields = ['created' => DateTimeFormat::utcNow()];
 253                                 $condition = ['nurl' => Strings::normaliseLink($server_url)];
 254                                 DBA::update('gserver', $fields, $condition);
 255                         }
 256
 257                         $last_contact = $gserver['last_contact'];
 258                         $last_failure = $gserver['last_failure'];
 259
 260                         // See discussion under https://forum.friendi.ca/display/0b6b25a8135aabc37a5a0f5684081633
 261                         // It can happen that a zero date is in the database, but storing it again is forbidden.
 262                         if ($last_contact < DBA::NULL_DATETIME) {
 263                                 $last_contact = DBA::NULL_DATETIME;
 264                         }
 265
 266                         if ($last_failure < DBA::NULL_DATETIME) {
 267                                 $last_failure = DBA::NULL_DATETIME;
 268                         }
 269
 270                         if (!$force && !self::updateNeeded($gserver['created'], '', $last_failure, $last_contact)) {
 271                                 Logger::info('No update needed', ['server' => $server_url]);
 272                                 return ($last_contact >= $last_failure);
 273                         }
 274                         Logger::info('Server is outdated. Start discovery.', ['Server' => $server_url, 'Force' => $force, 'Created' => $gserver['created'], 'Failure' => $last_failure, 'Contact' => $last_contact]);
 275                 } else {
 276                         Logger::info('Server is unknown. Start discovery.', ['Server' => $server_url]);
 277                 }
 278
 279                 return self::detect($server_url, $network, $only_nodeinfo);
 280         }
 281
 282         /**
 283          * Set failed server status
 284          *
 285          * @param string $url
 286          */
 287         public static function setFailure(string $url)
 288         {
 289                 $gserver = DBA::selectFirst('gserver', [], ['nurl' => Strings::normaliseLink($url)]);
 290                 if (DBA::isResult($gserver)) {
 291                         $next_update = self::getNextUpdateDate(false, $gserver['created'], $gserver['last_contact']);
 292                         DBA::update('gserver', ['failed' => true, 'last_failure' => DateTimeFormat::utcNow(),
 293                         'next_contact' => $next_update, 'detection-method' => null],
 294                         ['nurl' => Strings::normaliseLink($url)]);
 295                         Logger::info('Set failed status for existing server', ['url' => $url]);
 296                         return;
 297                 }
 298                 DBA::insert('gserver', ['url' => $url, 'nurl' => Strings::normaliseLink($url),
 299                         'network' => Protocol::PHANTOM, 'created' => DateTimeFormat::utcNow(),
 300                         'failed' => true, 'last_failure' => DateTimeFormat::utcNow()]);
 301                 Logger::info('Set failed status for new server', ['url' => $url]);
 302         }
 303
 304         /**
 305          * Remove unwanted content from the given URL
 306          *
 307          * @param string $url
 308          * @return string cleaned URL
 309          */
 310         public static function cleanURL(string $url)
 311         {
 312                 $url = trim($url, '/');
 313                 $url = str_replace('/index.php', '', $url);
 314
 315                 $urlparts = parse_url($url);
 316                 unset($urlparts['user']);
 317                 unset($urlparts['pass']);
 318                 unset($urlparts['query']);
 319                 unset($urlparts['fragment']);
 320                 return Network::unparseURL($urlparts);
 321         }
 322
 323         /**
 324          * Return the base URL
 325          *
 326          * @param string $url
 327          * @return string base URL
 328          */
 329         private static function getBaseURL(string $url)
 330         {
 331                 $urlparts = parse_url(self::cleanURL($url));
 332                 unset($urlparts['path']);
 333                 return Network::unparseURL($urlparts);
 334         }
 335
 336         /**
 337          * Detect server data (type, protocol, version number, ...)
 338          * The detected data is then updated or inserted in the gserver table.
 339          *
 340          * @param string  $url           URL of the given server
 341          * @param string  $network       Network value that is used, when detection failed
 342          * @param boolean $only_nodeinfo Only use nodeinfo for server detection
 343          *
 344          * @return boolean 'true' if server could be detected
 345          */
 346         public static function detect(string $url, string $network = '', bool $only_nodeinfo = false)
 347         {
 348                 Logger::info('Detect server type', ['server' => $url]);
 349                 $serverdata = ['detection-method' => self::DETECT_MANUAL];
 350
 351                 $original_url = $url;
 352
 353                 // Remove URL content that is not supposed to exist for a server url
 354                 $url = self::cleanURL($url);
 355
 356                 // Get base URL
 357                 $baseurl = self::getBaseURL($url);
 358
 359                 // If the URL missmatches, then we mark the old entry as failure
 360                 if ($url != $original_url) {
 361                         /// @todo What to do with "next_contact" here?
 362                         DBA::update('gserver', ['failed' => true, 'last_failure' => DateTimeFormat::utcNow()],
 363                                 ['nurl' => Strings::normaliseLink($original_url)]);
 364                 }
 365
 366                 // When a nodeinfo is present, we don't need to dig further
 367                 $xrd_timeout = DI::config()->get('system', 'xrd_timeout');
 368                 $curlResult = DI::httpRequest()->get($url . '/.well-known/nodeinfo', ['timeout' => $xrd_timeout]);
 369                 if ($curlResult->isTimeout()) {
 370                         self::setFailure($url);
 371                         return false;
 372                 }
 373
 374                 $nodeinfo = self::fetchNodeinfo($url, $curlResult);
 375                 if ($only_nodeinfo && empty($nodeinfo)) {
 376                         Logger::info('Invalid nodeinfo in nodeinfo-mode, server is marked as failure', ['url' => $url]);
 377                         self::setFailure($url);
 378                         return false;
 379                 }
 380
 381                 // When nodeinfo isn't present, we use the older 'statistics.json' endpoint
 382                 if (empty($nodeinfo)) {
 383                         $nodeinfo = self::fetchStatistics($url);
 384                 }
 385
 386                 // If that didn't work out well, we use some protocol specific endpoints
 387                 // For Friendica and Zot based networks we have to dive deeper to reveal more details
 388                 if (empty($nodeinfo['network']) || in_array($nodeinfo['network'], [Protocol::DFRN, Protocol::ZOT])) {
 389                         if (!empty($nodeinfo['detection-method'])) {
 390                                 $serverdata['detection-method'] = $nodeinfo['detection-method'];
 391                         }
 392
 393                         // Fetch the landing page, possibly it reveals some data
 394                         if (empty($nodeinfo['network'])) {
 395                                 if ($baseurl == $url) {
 396                                         $basedata = $serverdata;
 397                                 } else {
 398                                         $basedata = ['detection-method' => self::DETECT_MANUAL];
 399                                 }
 400
 401                                 $curlResult = DI::httpRequest()->get($baseurl, ['timeout' => $xrd_timeout]);
 402                                 if ($curlResult->isSuccess()) {
 403                                         $basedata = self::analyseRootHeader($curlResult, $basedata);
 404                                         $basedata = self::analyseRootBody($curlResult, $basedata, $baseurl);
 405                                 }
 406
 407                                 if (!$curlResult->isSuccess() || empty($curlResult->getBody()) || self::invalidBody($curlResult->getBody())) {
 408                                         self::setFailure($url);
 409                                         return false;
 410                                 }
 411
 412                                 if ($baseurl == $url) {
 413                                         $serverdata = $basedata;
 414                                 } else {
 415                                         // When the base path doesn't seem to contain a social network we try the complete path.
 416                                         // Most detectable system have to be installed in the root directory.
 417                                         // We checked the base to avoid false positives.
 418                                         $curlResult = DI::httpRequest()->get($url, ['timeout' => $xrd_timeout]);
 419                                         if ($curlResult->isSuccess()) {
 420                                                 $urldata = self::analyseRootHeader($curlResult, $serverdata);
 421                                                 $urldata = self::analyseRootBody($curlResult, $urldata, $url);
 422
 423                                                 $comparebase = $basedata;
 424                                                 unset($comparebase['info']);
 425                                                 unset($comparebase['site_name']);
 426                                                 $compareurl = $urldata;
 427                                                 unset($compareurl['info']);
 428                                                 unset($compareurl['site_name']);
 429
 430                                                 // We assume that no one will install the identical system in the root and a subfolder
 431                                                 if (!empty(array_diff($comparebase, $compareurl))) {
 432                                                         $serverdata = $urldata;
 433                                                 }
 434                                         }
 435                                 }
 436                         }
 437
 438                         if (empty($serverdata['network']) || ($serverdata['network'] == Protocol::ACTIVITYPUB)) {
 439                                 $serverdata = self::detectMastodonAlikes($url, $serverdata);
 440                         }
 441
 442                         // All following checks are done for systems that always have got a "host-meta" endpoint.
 443                         // With this check we don't have to waste time and ressources for dead systems.
 444                         // Also this hopefully prevents us from receiving abuse messages.
 445                         if (empty($serverdata['network']) && !self::validHostMeta($url)) {
 446                                 self::setFailure($url);
 447                                 return false;
 448                         }
 449
 450                         if (empty($serverdata['network']) || in_array($serverdata['network'], [Protocol::DFRN, Protocol::ACTIVITYPUB])) {
 451                                 $serverdata = self::detectFriendica($url, $serverdata);
 452                         }
 453
 454                         // the 'siteinfo.json' is some specific endpoint of Hubzilla and Red
 455                         if (empty($serverdata['network']) || ($serverdata['network'] == Protocol::ZOT)) {
 456                                 $serverdata = self::fetchSiteinfo($url, $serverdata);
 457                         }
 458
 459                         // The 'siteinfo.json' doesn't seem to be present on older Hubzilla installations
 460                         if (empty($serverdata['network'])) {
 461                                 $serverdata = self::detectHubzilla($url, $serverdata);
 462                         }
 463
 464                         if (empty($serverdata['network'])) {
 465                                 $serverdata = self::detectNextcloud($url, $serverdata);
 466                         }
 467
 468                         if (empty($serverdata['network'])) {
 469                                 $serverdata = self::detectGNUSocial($url, $serverdata);
 470                         }
 471
 472                         $serverdata = array_merge($nodeinfo, $serverdata);
 473                 } else {
 474                         $serverdata = $nodeinfo;
 475                 }
 476
 477                 // Detect the directory type
 478                 $serverdata['directory-type'] = self::DT_NONE;
 479                 $serverdata = self::checkPoCo($url, $serverdata);
 480                 $serverdata = self::checkMastodonDirectory($url, $serverdata);
 481
 482                 // We can't detect the network type. Possibly it is some system that we don't know yet
 483                 if (empty($serverdata['network'])) {
 484                         $serverdata['network'] = Protocol::PHANTOM;
 485                 }
 486
 487                 // When we hadn't been able to detect the network type, we use the hint from the parameter
 488                 if (($serverdata['network'] == Protocol::PHANTOM) && !empty($network)) {
 489                         $serverdata['network'] = $network;
 490                 }
 491
 492                 $serverdata['url'] = $url;
 493                 $serverdata['nurl'] = Strings::normaliseLink($url);
 494
 495                 // We take the highest number that we do find
 496                 $registeredUsers = $serverdata['registered-users'] ?? 0;
 497
 498                 // On an active server there has to be at least a single user
 499                 if (($serverdata['network'] != Protocol::PHANTOM) && ($registeredUsers == 0)) {
 500                         $registeredUsers = 1;
 501                 }
 502
 503                 if ($serverdata['network'] == Protocol::PHANTOM) {
 504                         $serverdata['registered-users'] = $registeredUsers;
 505                         $serverdata = self::detectNetworkViaContacts($url, $serverdata);
 506                 }
 507
 508                 $serverdata['next_contact'] = self::getNextUpdateDate(true);
 509
 510                 $serverdata['last_contact'] = DateTimeFormat::utcNow();
 511                 $serverdata['failed'] = false;
 512
 513                 $gserver = DBA::selectFirst('gserver', ['network'], ['nurl' => Strings::normaliseLink($url)]);
 514                 if (!DBA::isResult($gserver)) {
 515                         $serverdata['created'] = DateTimeFormat::utcNow();
 516                         $ret = DBA::insert('gserver', $serverdata);
 517                         $id = DBA::lastInsertId();
 518                 } else {
 519                         // Don't override the network with 'unknown' when there had been a valid entry before
 520                         if (($serverdata['network'] == Protocol::PHANTOM) && !empty($gserver['network'])) {
 521                                 unset($serverdata['network']);
 522                         }
 523
 524                         $ret = DBA::update('gserver', $serverdata, ['nurl' => $serverdata['nurl']]);
 525                         $gserver = DBA::selectFirst('gserver', ['id'], ['nurl' => $serverdata['nurl']]);
 526                         if (DBA::isResult($gserver)) {
 527                                 $id = $gserver['id'];
 528                         }
 529                 }
 530
 531                 if (!empty($serverdata['network']) && !empty($id) && ($serverdata['network'] != Protocol::PHANTOM)) {
 532                         $apcontacts = DBA::count('apcontact', ['gsid' => $id]);
 533                         $contacts = DBA::count('contact', ['uid' => 0, 'gsid' => $id]);
 534                         $max_users = max($apcontacts, $contacts, $registeredUsers);
 535                         if ($max_users > $registeredUsers) {
 536                                 Logger::info('Update registered users', ['id' => $id, 'url' => $serverdata['nurl'], 'registered-users' => $max_users]);
 537                                 DBA::update('gserver', ['registered-users' => $max_users], ['id' => $id]);
 538                         }
 539                 }
 540
 541                 if (!empty($serverdata['network']) && in_array($serverdata['network'], [Protocol::DFRN, Protocol::DIASPORA])) {
 542                         self::discoverRelay($url);
 543                 }
 544
 545                 return $ret;
 546         }
 547
 548         /**
 549          * Fetch relay data from a given server url
 550          *
 551          * @param string $server_url address of the server
 552          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 553          */
 554         private static function discoverRelay(string $server_url)
 555         {
 556                 Logger::info('Discover relay data', ['server' => $server_url]);
 557
 558                 $curlResult = DI::httpRequest()->get($server_url . '/.well-known/x-social-relay');
 559                 if (!$curlResult->isSuccess()) {
 560                         return;
 561                 }
 562
 563                 $data = json_decode($curlResult->getBody(), true);
 564                 if (!is_array($data)) {
 565                         return;
 566                 }
 567
 568                 // Sanitize incoming data, see https://github.com/friendica/friendica/issues/8565
 569                 $data['subscribe'] = (bool)$data['subscribe'] ?? false;
 570
 571                 if (!$data['subscribe'] || empty($data['scope']) || !in_array(strtolower($data['scope']), ['all', 'tags'])) {
 572                         $data['scope'] = '';
 573                         $data['subscribe'] = false;
 574                         $data['tags'] = [];
 575                 }
 576
 577                 $gserver = DBA::selectFirst('gserver', ['id', 'url', 'network', 'relay-subscribe', 'relay-scope'], ['nurl' => Strings::normaliseLink($server_url)]);
 578                 if (!DBA::isResult($gserver)) {
 579                         return;
 580                 }
 581
 582                 if (($gserver['relay-subscribe'] != $data['subscribe']) || ($gserver['relay-scope'] != $data['scope'])) {
 583                         $fields = ['relay-subscribe' => $data['subscribe'], 'relay-scope' => $data['scope']];
 584                         DBA::update('gserver', $fields, ['id' => $gserver['id']]);
 585                 }
 586
 587                 DBA::delete('gserver-tag', ['gserver-id' => $gserver['id']]);
 588
 589                 if ($data['scope'] == 'tags') {
 590                         // Avoid duplicates
 591                         $tags = [];
 592                         foreach ($data['tags'] as $tag) {
 593                                 $tag = mb_strtolower($tag);
 594                                 if (strlen($tag) < 100) {
 595                                         $tags[$tag] = $tag;
 596                                 }
 597                         }
 598
 599                         foreach ($tags as $tag) {
 600                                 DBA::insert('gserver-tag', ['gserver-id' => $gserver['id'], 'tag' => $tag], Database::INSERT_IGNORE);
 601                         }
 602                 }
 603
 604                 // Create or update the relay contact
 605                 $fields = [];
 606                 if (isset($data['protocols'])) {
 607                         if (isset($data['protocols']['diaspora'])) {
 608                                 $fields['network'] = Protocol::DIASPORA;
 609
 610                                 if (isset($data['protocols']['diaspora']['receive'])) {
 611                                         $fields['batch'] = $data['protocols']['diaspora']['receive'];
 612                                 } elseif (is_string($data['protocols']['diaspora'])) {
 613                                         $fields['batch'] = $data['protocols']['diaspora'];
 614                                 }
 615                         }
 616
 617                         if (isset($data['protocols']['dfrn'])) {
 618                                 $fields['network'] = Protocol::DFRN;
 619
 620                                 if (isset($data['protocols']['dfrn']['receive'])) {
 621                                         $fields['batch'] = $data['protocols']['dfrn']['receive'];
 622                                 } elseif (is_string($data['protocols']['dfrn'])) {
 623                                         $fields['batch'] = $data['protocols']['dfrn'];
 624                                 }
 625                         }
 626
 627                         if (isset($data['protocols']['activitypub'])) {
 628                                 $fields['network'] = Protocol::ACTIVITYPUB;
 629
 630                                 if (!empty($data['protocols']['activitypub']['actor'])) {
 631                                         $fields['url'] = $data['protocols']['activitypub']['actor'];
 632                                 }
 633                                 if (!empty($data['protocols']['activitypub']['receive'])) {
 634                                         $fields['batch'] = $data['protocols']['activitypub']['receive'];
 635                                 }
 636                         }
 637                 }
 638
 639                 Logger::info('Discovery ended', ['server' => $server_url, 'data' => $fields]);
 640
 641                 Relay::updateContact($gserver, $fields);
 642         }
 643
 644         /**
 645          * Fetch server data from '/statistics.json' on the given server
 646          *
 647          * @param string $url URL of the given server
 648          *
 649          * @return array server data
 650          */
 651         private static function fetchStatistics(string $url)
 652         {
 653                 $curlResult = DI::httpRequest()->get($url . '/statistics.json');
 654                 if (!$curlResult->isSuccess()) {
 655                         return [];
 656                 }
 657
 658                 $data = json_decode($curlResult->getBody(), true);
 659                 if (empty($data)) {
 660                         return [];
 661                 }
 662
 663                 $serverdata = ['detection-method' => self::DETECT_STATISTICS_JSON];
 664
 665                 if (!empty($data['version'])) {
 666                         $serverdata['version'] = $data['version'];
 667                         // Version numbers on statistics.json are presented with additional info, e.g.:
 668                         // 0.6.3.0-p1702cc1c, 0.6.99.0-p1b9ab160 or 3.4.3-2-1191.
 669                         $serverdata['version'] = preg_replace('=(.+)-(.{4,})=ism', '$1', $serverdata['version']);
 670                 }
 671
 672                 if (!empty($data['name'])) {
 673                         $serverdata['site_name'] = $data['name'];
 674                 }
 675
 676                 if (!empty($data['network'])) {
 677                         $serverdata['platform'] = strtolower($data['network']);
 678
 679                         if ($serverdata['platform'] == 'diaspora') {
 680                                 $serverdata['network'] = Protocol::DIASPORA;
 681                         } elseif ($serverdata['platform'] == 'friendica') {
 682                                 $serverdata['network'] = Protocol::DFRN;
 683                         } elseif ($serverdata['platform'] == 'hubzilla') {
 684                                 $serverdata['network'] = Protocol::ZOT;
 685                         } elseif ($serverdata['platform'] == 'redmatrix') {
 686                                 $serverdata['network'] = Protocol::ZOT;
 687                         }
 688                 }
 689
 690
 691                 if (!empty($data['registrations_open'])) {
 692                         $serverdata['register_policy'] = Register::OPEN;
 693                 } else {
 694                         $serverdata['register_policy'] = Register::CLOSED;
 695                 }
 696
 697                 return $serverdata;
 698         }
 699
 700         /**
 701          * Detect server type by using the nodeinfo data
 702          *
 703          * @param string     $url        address of the server
 704          * @param CurlResult $curlResult
 705          * @return array Server data
 706          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 707          */
 708         private static function fetchNodeinfo(string $url, CurlResult $curlResult)
 709         {
 710                 if (!$curlResult->isSuccess()) {
 711                         return [];
 712                 }
 713
 714                 $nodeinfo = json_decode($curlResult->getBody(), true);
 715
 716                 if (!is_array($nodeinfo) || empty($nodeinfo['links'])) {
 717                         return [];
 718                 }
 719
 720                 $nodeinfo1_url = '';
 721                 $nodeinfo2_url = '';
 722
 723                 foreach ($nodeinfo['links'] as $link) {
 724                         if (!is_array($link) || empty($link['rel']) || empty($link['href'])) {
 725                                 Logger::info('Invalid nodeinfo format', ['url' => $url]);
 726                                 continue;
 727                         }
 728                         if ($link['rel'] == 'http://nodeinfo.diaspora.software/ns/schema/1.0') {
 729                                 $nodeinfo1_url = $link['href'];
 730                         } elseif ($link['rel'] == 'http://nodeinfo.diaspora.software/ns/schema/2.0') {
 731                                 $nodeinfo2_url = $link['href'];
 732                         }
 733                 }
 734
 735                 if ($nodeinfo1_url . $nodeinfo2_url == '') {
 736                         return [];
 737                 }
 738
 739                 $server = [];
 740
 741                 // When the nodeinfo url isn't on the same host, then there is obviously something wrong
 742                 if (!empty($nodeinfo2_url) && (parse_url($url, PHP_URL_HOST) == parse_url($nodeinfo2_url, PHP_URL_HOST))) {
 743                         $server = self::parseNodeinfo2($nodeinfo2_url);
 744                 }
 745
 746                 // When the nodeinfo url isn't on the same host, then there is obviously something wrong
 747                 if (empty($server) && !empty($nodeinfo1_url) && (parse_url($url, PHP_URL_HOST) == parse_url($nodeinfo1_url, PHP_URL_HOST))) {
 748                         $server = self::parseNodeinfo1($nodeinfo1_url);
 749                 }
 750
 751                 return $server;
 752         }
 753
 754         /**
 755          * Parses Nodeinfo 1
 756          *
 757          * @param string $nodeinfo_url address of the nodeinfo path
 758          * @return array Server data
 759          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 760          */
 761         private static function parseNodeinfo1(string $nodeinfo_url)
 762         {
 763                 $curlResult = DI::httpRequest()->get($nodeinfo_url);
 764
 765                 if (!$curlResult->isSuccess()) {
 766                         return [];
 767                 }
 768
 769                 $nodeinfo = json_decode($curlResult->getBody(), true);
 770
 771                 if (!is_array($nodeinfo)) {
 772                         return [];
 773                 }
 774
 775                 $server = ['detection-method' => self::DETECT_NODEINFO_1,
 776                         'register_policy' => Register::CLOSED];
 777
 778                 if (!empty($nodeinfo['openRegistrations'])) {
 779                         $server['register_policy'] = Register::OPEN;
 780                 }
 781
 782                 if (is_array($nodeinfo['software'])) {
 783                         if (!empty($nodeinfo['software']['name'])) {
 784                                 $server['platform'] = strtolower($nodeinfo['software']['name']);
 785                         }
 786
 787                         if (!empty($nodeinfo['software']['version'])) {
 788                                 $server['version'] = $nodeinfo['software']['version'];
 789                                 // Version numbers on Nodeinfo are presented with additional info, e.g.:
 790                                 // 0.6.3.0-p1702cc1c, 0.6.99.0-p1b9ab160 or 3.4.3-2-1191.
 791                                 $server['version'] = preg_replace('=(.+)-(.{4,})=ism', '$1', $server['version']);
 792                         }
 793                 }
 794
 795                 if (!empty($nodeinfo['metadata']['nodeName'])) {
 796                         $server['site_name'] = $nodeinfo['metadata']['nodeName'];
 797                 }
 798
 799                 if (!empty($nodeinfo['usage']['users']['total'])) {
 800                         $server['registered-users'] = $nodeinfo['usage']['users']['total'];
 801                 }
 802
 803                 if (!empty($nodeinfo['protocols']['inbound']) && is_array($nodeinfo['protocols']['inbound'])) {
 804                         $protocols = [];
 805                         foreach ($nodeinfo['protocols']['inbound'] as $protocol) {
 806                                 $protocols[$protocol] = true;
 807                         }
 808
 809                         if (!empty($protocols['friendica'])) {
 810                                 $server['network'] = Protocol::DFRN;
 811                         } elseif (!empty($protocols['activitypub'])) {
 812                                 $server['network'] = Protocol::ACTIVITYPUB;
 813                         } elseif (!empty($protocols['diaspora'])) {
 814                                 $server['network'] = Protocol::DIASPORA;
 815                         } elseif (!empty($protocols['ostatus'])) {
 816                                 $server['network'] = Protocol::OSTATUS;
 817                         } elseif (!empty($protocols['gnusocial'])) {
 818                                 $server['network'] = Protocol::OSTATUS;
 819                         } elseif (!empty($protocols['zot'])) {
 820                                 $server['network'] = Protocol::ZOT;
 821                         }
 822                 }
 823
 824                 if (empty($server)) {
 825                         return [];
 826                 }
 827
 828                 return $server;
 829         }
 830
 831         /**
 832          * Parses Nodeinfo 2
 833          *
 834          * @param string $nodeinfo_url address of the nodeinfo path
 835          * @return array Server data
 836          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 837          */
 838         private static function parseNodeinfo2(string $nodeinfo_url)
 839         {
 840                 $curlResult = DI::httpRequest()->get($nodeinfo_url);
 841                 if (!$curlResult->isSuccess()) {
 842                         return [];
 843                 }
 844
 845                 $nodeinfo = json_decode($curlResult->getBody(), true);
 846
 847                 if (!is_array($nodeinfo)) {
 848                         return [];
 849                 }
 850
 851                 $server = ['detection-method' => self::DETECT_NODEINFO_2,
 852                         'register_policy' => Register::CLOSED];
 853
 854                 if (!empty($nodeinfo['openRegistrations'])) {
 855                         $server['register_policy'] = Register::OPEN;
 856                 }
 857
 858                 if (is_array($nodeinfo['software'])) {
 859                         if (!empty($nodeinfo['software']['name'])) {
 860                                 $server['platform'] = strtolower($nodeinfo['software']['name']);
 861                         }
 862
 863                         if (!empty($nodeinfo['software']['version'])) {
 864                                 $server['version'] = $nodeinfo['software']['version'];
 865                                 // Version numbers on Nodeinfo are presented with additional info, e.g.:
 866                                 // 0.6.3.0-p1702cc1c, 0.6.99.0-p1b9ab160 or 3.4.3-2-1191.
 867                                 $server['version'] = preg_replace('=(.+)-(.{4,})=ism', '$1', $server['version']);
 868                         }
 869                 }
 870
 871                 if (!empty($nodeinfo['metadata']['nodeName'])) {
 872                         $server['site_name'] = $nodeinfo['metadata']['nodeName'];
 873                 }
 874
 875                 if (!empty($nodeinfo['usage']['users']['total'])) {
 876                         $server['registered-users'] = $nodeinfo['usage']['users']['total'];
 877                 }
 878
 879                 if (!empty($nodeinfo['protocols'])) {
 880                         $protocols = [];
 881                         foreach ($nodeinfo['protocols'] as $protocol) {
 882                                 $protocols[$protocol] = true;
 883                         }
 884
 885                         if (!empty($protocols['dfrn'])) {
 886                                 $server['network'] = Protocol::DFRN;
 887                         } elseif (!empty($protocols['activitypub'])) {
 888                                 $server['network'] = Protocol::ACTIVITYPUB;
 889                         } elseif (!empty($protocols['diaspora'])) {
 890                                 $server['network'] = Protocol::DIASPORA;
 891                         } elseif (!empty($protocols['ostatus'])) {
 892                                 $server['network'] = Protocol::OSTATUS;
 893                         } elseif (!empty($protocols['gnusocial'])) {
 894                                 $server['network'] = Protocol::OSTATUS;
 895                         } elseif (!empty($protocols['zot'])) {
 896                                 $server['network'] = Protocol::ZOT;
 897                         }
 898                 }
 899
 900                 if (empty($server)) {
 901                         return [];
 902                 }
 903
 904                 return $server;
 905         }
 906
 907         /**
 908          * Fetch server information from a 'siteinfo.json' file on the given server
 909          *
 910          * @param string $url        URL of the given server
 911          * @param array  $serverdata array with server data
 912          *
 913          * @return array server data
 914          */
 915         private static function fetchSiteinfo(string $url, array $serverdata)
 916         {
 917                 $curlResult = DI::httpRequest()->get($url . '/siteinfo.json');
 918                 if (!$curlResult->isSuccess()) {
 919                         return $serverdata;
 920                 }
 921
 922                 $data = json_decode($curlResult->getBody(), true);
 923                 if (empty($data)) {
 924                         return $serverdata;
 925                 }
 926
 927                 if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
 928                         $serverdata['detection-method'] = self::DETECT_SITEINFO_JSON;
 929                 }
 930
 931                 if (!empty($data['url'])) {
 932                         $serverdata['platform'] = strtolower($data['platform']);
 933                         $serverdata['version'] = $data['version'];
 934                 }
 935
 936                 if (!empty($data['plugins'])) {
 937                         if (in_array('pubcrawl', $data['plugins'])) {
 938                                 $serverdata['network'] = Protocol::ACTIVITYPUB;
 939                         } elseif (in_array('diaspora', $data['plugins'])) {
 940                                 $serverdata['network'] = Protocol::DIASPORA;
 941                         } elseif (in_array('gnusoc', $data['plugins'])) {
 942                                 $serverdata['network'] = Protocol::OSTATUS;
 943                         } else {
 944                                 $serverdata['network'] = Protocol::ZOT;
 945                         }
 946                 }
 947
 948                 if (!empty($data['site_name'])) {
 949                         $serverdata['site_name'] = $data['site_name'];
 950                 }
 951
 952                 if (!empty($data['channels_total'])) {
 953                         $serverdata['registered-users'] = $data['channels_total'];
 954                 }
 955
 956                 if (!empty($data['register_policy'])) {
 957                         switch ($data['register_policy']) {
 958                                 case 'REGISTER_OPEN':
 959                                         $serverdata['register_policy'] = Register::OPEN;
 960                                         break;
 961
 962                                 case 'REGISTER_APPROVE':
 963                                         $serverdata['register_policy'] = Register::APPROVE;
 964                                         break;
 965
 966                                 case 'REGISTER_CLOSED':
 967                                 default:
 968                                         $serverdata['register_policy'] = Register::CLOSED;
 969                                         break;
 970                         }
 971                 }
 972
 973                 return $serverdata;
 974         }
 975
 976         /**
 977          * Checks if the server contains a valid host meta file
 978          *
 979          * @param string $url URL of the given server
 980          *
 981          * @return boolean 'true' if the server seems to be vital
 982          */
 983         private static function validHostMeta(string $url)
 984         {
 985                 $xrd_timeout = DI::config()->get('system', 'xrd_timeout');
 986                 $curlResult = DI::httpRequest()->get($url . '/.well-known/host-meta', ['timeout' => $xrd_timeout]);
 987                 if (!$curlResult->isSuccess()) {
 988                         return false;
 989                 }
 990
 991                 $xrd = XML::parseString($curlResult->getBody());
 992                 if (!is_object($xrd)) {
 993                         return false;
 994                 }
 995
 996                 $elements = XML::elementToArray($xrd);
 997                 if (empty($elements) || empty($elements['xrd']) || empty($elements['xrd']['link'])) {
 998                         return false;
 999                 }
1000
1001                 $valid = false;
1002                 foreach ($elements['xrd']['link'] as $link) {
1003                         // When there is more than a single "link" element, the array looks slightly different
1004                         if (!empty($link['@attributes'])) {
1005                                 $link = $link['@attributes'];
1006                         }
1007
1008                         if (empty($link['rel']) || empty($link['template'])) {
1009                                 continue;
1010                         }
1011
1012                         if ($link['rel'] == 'lrdd') {
1013                                 // When the webfinger host is the same like the system host, it should be ok.
1014                                 $valid = (parse_url($url, PHP_URL_HOST) == parse_url($link['template'], PHP_URL_HOST));
1015                         }
1016                 }
1017
1018                 return $valid;
1019         }
1020
1021         /**
1022          * Detect the network of the given server via their known contacts
1023          *
1024          * @param string $url        URL of the given server
1025          * @param array  $serverdata array with server data
1026          *
1027          * @return array server data
1028          */
1029         private static function detectNetworkViaContacts(string $url, array $serverdata)
1030         {
1031                 $contacts = [];
1032
1033                 $apcontacts = DBA::select('apcontact', ['url'], ['baseurl' => [$url, $serverdata['nurl']]]);
1034                 while ($apcontact = DBA::fetch($apcontacts)) {
1035                         $contacts[Strings::normaliseLink($apcontact['url'])] = $apcontact['url'];
1036                 }
1037                 DBA::close($apcontacts);
1038
1039                 $pcontacts = DBA::select('contact', ['url', 'nurl'], ['uid' => 0, 'baseurl' => [$url, $serverdata['nurl']]]);
1040                 while ($pcontact = DBA::fetch($pcontacts)) {
1041                         $contacts[$pcontact['nurl']] = $pcontact['url'];
1042                 }
1043                 DBA::close($pcontacts);
1044
1045                 if (empty($contacts)) {
1046                         return $serverdata;
1047                 }
1048
1049                 foreach ($contacts as $contact) {
1050                         $probed = Contact::getByURL($contact);
1051                         if (!empty($probed) && in_array($probed['network'], Protocol::FEDERATED)) {
1052                                 $serverdata['network'] = $probed['network'];
1053                                 break;
1054                         }
1055                 }
1056
1057                 $serverdata['registered-users'] = max($serverdata['registered-users'], count($contacts));
1058
1059                 return $serverdata;
1060         }
1061
1062         /**
1063          * Checks if the given server does have a '/poco' endpoint.
1064          * This is used for the 'PortableContact' functionality,
1065          * which is used by both Friendica and Hubzilla.
1066          *
1067          * @param string $url        URL of the given server
1068          * @param array  $serverdata array with server data
1069          *
1070          * @return array server data
1071          */
1072         private static function checkPoCo(string $url, array $serverdata)
1073         {
1074                 $serverdata['poco'] = '';
1075
1076                 $curlResult = DI::httpRequest()->get($url . '/poco');
1077                 if (!$curlResult->isSuccess()) {
1078                         return $serverdata;
1079                 }
1080
1081                 $data = json_decode($curlResult->getBody(), true);
1082                 if (empty($data)) {
1083                         return $serverdata;
1084                 }
1085
1086                 if (!empty($data['totalResults'])) {
1087                         $registeredUsers = $serverdata['registered-users'] ?? 0;
1088                         $serverdata['registered-users'] = max($data['totalResults'], $registeredUsers);
1089                         $serverdata['directory-type'] = self::DT_POCO;
1090                         $serverdata['poco'] = $url . '/poco';
1091                 }
1092
1093                 return $serverdata;
1094         }
1095
1096         /**
1097          * Checks if the given server does have a Mastodon style directory endpoint.
1098          *
1099          * @param string $url        URL of the given server
1100          * @param array  $serverdata array with server data
1101          *
1102          * @return array server data
1103          */
1104         public static function checkMastodonDirectory(string $url, array $serverdata)
1105         {
1106                 $curlResult = DI::httpRequest()->get($url . '/api/v1/directory?limit=1');
1107                 if (!$curlResult->isSuccess()) {
1108                         return $serverdata;
1109                 }
1110
1111                 $data = json_decode($curlResult->getBody(), true);
1112                 if (empty($data)) {
1113                         return $serverdata;
1114                 }
1115
1116                 if (count($data) == 1) {
1117                         $serverdata['directory-type'] = self::DT_MASTODON;
1118                 }
1119
1120                 return $serverdata;
1121         }
1122
1123         /**
1124          * Detects the version number of a given server when it was a NextCloud installation
1125          *
1126          * @param string $url        URL of the given server
1127          * @param array  $serverdata array with server data
1128          *
1129          * @return array server data
1130          */
1131         private static function detectNextcloud(string $url, array $serverdata)
1132         {
1133                 $curlResult = DI::httpRequest()->get($url . '/status.php');
1134
1135                 if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) {
1136                         return $serverdata;
1137                 }
1138
1139                 $data = json_decode($curlResult->getBody(), true);
1140                 if (empty($data)) {
1141                         return $serverdata;
1142                 }
1143
1144                 if (!empty($data['version'])) {
1145                         $serverdata['platform'] = 'nextcloud';
1146                         $serverdata['version'] = $data['version'];
1147                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1148
1149                         if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1150                                 $serverdata['detection-method'] = self::DETECT_STATUS_PHP;
1151                         }
1152                 }
1153
1154                 return $serverdata;
1155         }
1156
1157         /**
1158          * Detects data from a given server url if it was a mastodon alike system
1159          *
1160          * @param string $url        URL of the given server
1161          * @param array  $serverdata array with server data
1162          *
1163          * @return array server data
1164          */
1165         private static function detectMastodonAlikes(string $url, array $serverdata)
1166         {
1167                 $curlResult = DI::httpRequest()->get($url . '/api/v1/instance');
1168
1169                 if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) {
1170                         return $serverdata;
1171                 }
1172
1173                 $data = json_decode($curlResult->getBody(), true);
1174                 if (empty($data)) {
1175                         return $serverdata;
1176                 }
1177
1178                 if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1179                         $serverdata['detection-method'] = self::DETECT_MASTODON_API;
1180                 }
1181
1182                 if (!empty($data['version'])) {
1183                         $serverdata['platform'] = 'mastodon';
1184                         $serverdata['version'] = $data['version'] ?? '';
1185                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1186                 }
1187
1188                 if (!empty($data['title'])) {
1189                         $serverdata['site_name'] = $data['title'];
1190                 }
1191
1192                 if (!empty($data['title']) && empty($serverdata['platform']) && empty($serverdata['network'])) {
1193                         $serverdata['platform'] = 'mastodon';
1194                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1195                 }
1196
1197                 if (!empty($data['description'])) {
1198                         $serverdata['info'] = trim($data['description']);
1199                 }
1200
1201                 if (!empty($data['stats']['user_count'])) {
1202                         $serverdata['registered-users'] = $data['stats']['user_count'];
1203                 }
1204
1205                 if (!empty($serverdata['version']) && preg_match('/.*?\(compatible;\s(.*)\s(.*)\)/ism', $serverdata['version'], $matches)) {
1206                         $serverdata['platform'] = strtolower($matches[1]);
1207                         $serverdata['version'] = $matches[2];
1208                 }
1209
1210                 if (!empty($serverdata['version']) && strstr(strtolower($serverdata['version']), 'pleroma')) {
1211                         $serverdata['platform'] = 'pleroma';
1212                         $serverdata['version'] = trim(str_ireplace('pleroma', '', $serverdata['version']));
1213                 }
1214
1215                 if (!empty($serverdata['platform']) && strstr($serverdata['platform'], 'pleroma')) {
1216                         $serverdata['version'] = trim(str_ireplace('pleroma', '', $serverdata['platform']));
1217                         $serverdata['platform'] = 'pleroma';
1218                 }
1219
1220                 return $serverdata;
1221         }
1222
1223         /**
1224          * Detects data from typical Hubzilla endpoints
1225          *
1226          * @param string $url        URL of the given server
1227          * @param array  $serverdata array with server data
1228          *
1229          * @return array server data
1230          */
1231         private static function detectHubzilla(string $url, array $serverdata)
1232         {
1233                 $curlResult = DI::httpRequest()->get($url . '/api/statusnet/config.json');
1234                 if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) {
1235                         return $serverdata;
1236                 }
1237
1238                 $data = json_decode($curlResult->getBody(), true);
1239                 if (empty($data) || empty($data['site'])) {
1240                         return $serverdata;
1241                 }
1242
1243                 if (!empty($data['site']['name'])) {
1244                         $serverdata['site_name'] = $data['site']['name'];
1245                 }
1246
1247                 if (!empty($data['site']['platform'])) {
1248                         $serverdata['platform'] = strtolower($data['site']['platform']['PLATFORM_NAME']);
1249                         $serverdata['version'] = $data['site']['platform']['STD_VERSION'];
1250                         $serverdata['network'] = Protocol::ZOT;
1251                 }
1252
1253                 if (!empty($data['site']['hubzilla'])) {
1254                         $serverdata['platform'] = strtolower($data['site']['hubzilla']['PLATFORM_NAME']);
1255                         $serverdata['version'] = $data['site']['hubzilla']['RED_VERSION'];
1256                         $serverdata['network'] = Protocol::ZOT;
1257                 }
1258
1259                 if (!empty($data['site']['redmatrix'])) {
1260                         if (!empty($data['site']['redmatrix']['PLATFORM_NAME'])) {
1261                                 $serverdata['platform'] = strtolower($data['site']['redmatrix']['PLATFORM_NAME']);
1262                         } elseif (!empty($data['site']['redmatrix']['RED_PLATFORM'])) {
1263                                 $serverdata['platform'] = strtolower($data['site']['redmatrix']['RED_PLATFORM']);
1264                         }
1265
1266                         $serverdata['version'] = $data['site']['redmatrix']['RED_VERSION'];
1267                         $serverdata['network'] = Protocol::ZOT;
1268                 }
1269
1270                 $private = false;
1271                 $inviteonly = false;
1272                 $closed = false;
1273
1274                 if (!empty($data['site']['closed'])) {
1275                         $closed = self::toBoolean($data['site']['closed']);
1276                 }
1277
1278                 if (!empty($data['site']['private'])) {
1279                         $private = self::toBoolean($data['site']['private']);
1280                 }
1281
1282                 if (!empty($data['site']['inviteonly'])) {
1283                         $inviteonly = self::toBoolean($data['site']['inviteonly']);
1284                 }
1285
1286                 if (!$closed && !$private and $inviteonly) {
1287                         $serverdata['register_policy'] = Register::APPROVE;
1288                 } elseif (!$closed && !$private) {
1289                         $serverdata['register_policy'] = Register::OPEN;
1290                 } else {
1291                         $serverdata['register_policy'] = Register::CLOSED;
1292                 }
1293
1294                 if (!empty($serverdata['network']) && in_array($serverdata['detection-method'],
1295                         [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1296                         $serverdata['detection-method'] = self::DETECT_CONFIG_JSON;
1297                 }
1298
1299                 return $serverdata;
1300         }
1301
1302         /**
1303          * Converts input value to a boolean value
1304          *
1305          * @param string|integer $val
1306          *
1307          * @return boolean
1308          */
1309         private static function toBoolean($val)
1310         {
1311                 if (($val == 'true') || ($val == 1)) {
1312                         return true;
1313                 } elseif (($val == 'false') || ($val == 0)) {
1314                         return false;
1315                 }
1316
1317                 return $val;
1318         }
1319
1320         /**
1321          * Detect if the URL belongs to a GNU Social server
1322          *
1323          * @param string $url        URL of the given server
1324          * @param array  $serverdata array with server data
1325          *
1326          * @return array server data
1327          */
1328         private static function detectGNUSocial(string $url, array $serverdata)
1329         {
1330                 // Test for GNU Social
1331                 $curlResult = DI::httpRequest()->get($url . '/api/gnusocial/version.json');
1332                 if ($curlResult->isSuccess() && ($curlResult->getBody() != '{"error":"not implemented"}') &&
1333                         ($curlResult->getBody() != '') && (strlen($curlResult->getBody()) < 30)) {
1334                         $serverdata['platform'] = 'gnusocial';
1335                         // Remove junk that some GNU Social servers return
1336                         $serverdata['version'] = str_replace(chr(239) . chr(187) . chr(191), '', $curlResult->getBody());
1337                         $serverdata['version'] = str_replace(["\r", "\n", "\t"], '', $serverdata['version']);
1338                         $serverdata['version'] = trim($serverdata['version'], '"');
1339                         $serverdata['network'] = Protocol::OSTATUS;
1340
1341                         if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1342                                 $serverdata['detection-method'] = self::DETECT_GNUSOCIAL;
1343                         }
1344
1345                         return $serverdata;
1346                 }
1347
1348                 // Test for Statusnet
1349                 $curlResult = DI::httpRequest()->get($url . '/api/statusnet/version.json');
1350                 if ($curlResult->isSuccess() && ($curlResult->getBody() != '{"error":"not implemented"}') &&
1351                         ($curlResult->getBody() != '') && (strlen($curlResult->getBody()) < 30)) {
1352
1353                         // Remove junk that some GNU Social servers return
1354                         $serverdata['version'] = str_replace(chr(239).chr(187).chr(191), '', $curlResult->getBody());
1355                         $serverdata['version'] = str_replace(["\r", "\n", "\t"], '', $serverdata['version']);
1356                         $serverdata['version'] = trim($serverdata['version'], '"');
1357
1358                         if (!empty($serverdata['version']) && strtolower(substr($serverdata['version'], 0, 7)) == 'pleroma') {
1359                                 $serverdata['platform'] = 'pleroma';
1360                                 $serverdata['version'] = trim(str_ireplace('pleroma', '', $serverdata['version']));
1361                                 $serverdata['network'] = Protocol::ACTIVITYPUB;
1362                         } else {
1363                                 $serverdata['platform'] = 'statusnet';
1364                                 $serverdata['network'] = Protocol::OSTATUS;
1365                         }
1366
1367                         if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1368                                 $serverdata['detection-method'] = self::DETECT_STATUSNET;
1369                         }
1370                 }
1371
1372                 return $serverdata;
1373         }
1374
1375         /**
1376          * Detect if the URL belongs to a Friendica server
1377          *
1378          * @param string $url        URL of the given server
1379          * @param array  $serverdata array with server data
1380          *
1381          * @return array server data
1382          */
1383         private static function detectFriendica(string $url, array $serverdata)
1384         {
1385                 $curlResult = DI::httpRequest()->get($url . '/friendica/json');
1386                 if (!$curlResult->isSuccess()) {
1387                         $curlResult = DI::httpRequest()->get($url . '/friendika/json');
1388                         $friendika = true;
1389                         $platform = 'Friendika';
1390                 } else {
1391                         $friendika = false;
1392                         $platform = 'Friendica';
1393                 }
1394
1395                 if (!$curlResult->isSuccess()) {
1396                         return $serverdata;
1397                 }
1398
1399                 $data = json_decode($curlResult->getBody(), true);
1400                 if (empty($data) || empty($data['version'])) {
1401                         return $serverdata;
1402                 }
1403
1404                 if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) {
1405                         $serverdata['detection-method'] = $friendika ? self::DETECT_FRIENDIKA : self::DETECT_FRIENDICA;
1406                 }
1407
1408                 $serverdata['network'] = Protocol::DFRN;
1409                 $serverdata['version'] = $data['version'];
1410
1411                 if (!empty($data['no_scrape_url'])) {
1412                         $serverdata['noscrape'] = $data['no_scrape_url'];
1413                 }
1414
1415                 if (!empty($data['site_name'])) {
1416                         $serverdata['site_name'] = $data['site_name'];
1417                 }
1418
1419                 if (!empty($data['info'])) {
1420                         $serverdata['info'] = trim($data['info']);
1421                 }
1422
1423                 $register_policy = ($data['register_policy'] ?? '') ?: 'REGISTER_CLOSED';
1424                 switch ($register_policy) {
1425                         case 'REGISTER_OPEN':
1426                                 $serverdata['register_policy'] = Register::OPEN;
1427                                 break;
1428
1429                         case 'REGISTER_APPROVE':
1430                                 $serverdata['register_policy'] = Register::APPROVE;
1431                                 break;
1432
1433                         case 'REGISTER_CLOSED':
1434                         case 'REGISTER_INVITATION':
1435                                 $serverdata['register_policy'] = Register::CLOSED;
1436                                 break;
1437                         default:
1438                                 Logger::info('Register policy is invalid', ['policy' => $register_policy, 'server' => $url]);
1439                                 $serverdata['register_policy'] = Register::CLOSED;
1440                                 break;
1441                 }
1442
1443                 $serverdata['platform'] = strtolower($data['platform'] ?? $platform);
1444
1445                 return $serverdata;
1446         }
1447
1448         /**
1449          * Analyses the landing page of a given server for hints about type and system of that server
1450          *
1451          * @param object $curlResult result of curl execution
1452          * @param array  $serverdata array with server data
1453          * @param string $url        Server URL
1454          *
1455          * @return array server data
1456          */
1457         private static function analyseRootBody($curlResult, array $serverdata, string $url)
1458         {
1459                 $doc = new DOMDocument();
1460                 @$doc->loadHTML($curlResult->getBody());
1461                 $xpath = new DOMXPath($doc);
1462
1463                 $title = trim(XML::getFirstNodeValue($xpath, '//head/title/text()'));
1464                 if (!empty($title)) {
1465                         $serverdata['site_name'] = $title;
1466                 }
1467
1468                 $list = $xpath->query('//meta[@name]');
1469
1470                 foreach ($list as $node) {
1471                         $attr = [];
1472                         if ($node->attributes->length) {
1473                                 foreach ($node->attributes as $attribute) {
1474                                         $value = trim($attribute->value);
1475                                         if (empty($value)) {
1476                                                 continue;
1477                                         }
1478
1479                                         $attr[$attribute->name] = $value;
1480                                 }
1481
1482                                 if (empty($attr['name']) || empty($attr['content'])) {
1483                                         continue;
1484                                 }
1485                         }
1486
1487                         if ($attr['name'] == 'description') {
1488                                 $serverdata['info'] = $attr['content'];
1489                         }
1490
1491                         if (in_array($attr['name'], ['application-name', 'al:android:app_name', 'al:ios:app_name',
1492                                 'twitter:app:name:googleplay', 'twitter:app:name:iphone', 'twitter:app:name:ipad'])) {
1493                                 $serverdata['platform'] = strtolower($attr['content']);
1494                                 if (in_array($attr['content'], ['Misskey', 'Write.as'])) {
1495                                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1496                                 }
1497                         }
1498                         if (($attr['name'] == 'generator') && (empty($serverdata['platform']) || (substr(strtolower($attr['content']), 0, 9) == 'wordpress'))) {
1499                                 $serverdata['platform'] = strtolower($attr['content']);
1500                                 $version_part = explode(' ', $attr['content']);
1501
1502                                 if (count($version_part) == 2) {
1503                                         if (in_array($version_part[0], ['WordPress'])) {
1504                                                 $serverdata['platform'] = strtolower($version_part[0]);
1505                                                 $serverdata['version'] = $version_part[1];
1506
1507                                                 // We still do need a reliable test if some AP plugin is activated
1508                                                 if (DBA::exists('apcontact', ['baseurl' => $url])) {
1509                                                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1510                                                 } else {
1511                                                         $serverdata['network'] = Protocol::FEED;
1512                                                 }
1513
1514                                                 if ($serverdata['detection-method'] == self::DETECT_MANUAL) {
1515                                                         $serverdata['detection-method'] = self::DETECT_BODY;
1516                                                 }
1517                                         }
1518                                         if (in_array($version_part[0], ['Friendika', 'Friendica'])) {
1519                                                 $serverdata['platform'] = strtolower($version_part[0]);
1520                                                 $serverdata['version'] = $version_part[1];
1521                                                 $serverdata['network'] = Protocol::DFRN;
1522                                         }
1523                                 }
1524                         }
1525                 }
1526
1527                 $list = $xpath->query('//meta[@property]');
1528
1529                 foreach ($list as $node) {
1530                         $attr = [];
1531                         if ($node->attributes->length) {
1532                                 foreach ($node->attributes as $attribute) {
1533                                         $value = trim($attribute->value);
1534                                         if (empty($value)) {
1535                                                 continue;
1536                                         }
1537
1538                                         $attr[$attribute->name] = $value;
1539                                 }
1540
1541                                 if (empty($attr['property']) || empty($attr['content'])) {
1542                                         continue;
1543                                 }
1544                         }
1545
1546                         if ($attr['property'] == 'og:site_name') {
1547                                 $serverdata['site_name'] = $attr['content'];
1548                         }
1549
1550                         if ($attr['property'] == 'og:description') {
1551                                 $serverdata['info'] = $attr['content'];
1552                         }
1553
1554                         if ($attr['property'] == 'og:platform') {
1555                                 $serverdata['platform'] = strtolower($attr['content']);
1556
1557                                 if (in_array($attr['content'], ['PeerTube'])) {
1558                                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1559                                 }
1560                         }
1561
1562                         if ($attr['property'] == 'generator') {
1563                                 $serverdata['platform'] = strtolower($attr['content']);
1564
1565                                 if (in_array($attr['content'], ['hubzilla'])) {
1566                                         // We later check which compatible protocol modules are loaded.
1567                                         $serverdata['network'] = Protocol::ZOT;
1568                                 }
1569                         }
1570                 }
1571
1572                 if (!empty($serverdata['network']) && ($serverdata['detection-method'] == self::DETECT_MANUAL)) {
1573                         $serverdata['detection-method'] = self::DETECT_BODY;
1574                 }
1575
1576                 return $serverdata;
1577         }
1578
1579         /**
1580          * Analyses the header data of a given server for hints about type and system of that server
1581          *
1582          * @param object $curlResult result of curl execution
1583          * @param array  $serverdata array with server data
1584          *
1585          * @return array server data
1586          */
1587         private static function analyseRootHeader($curlResult, array $serverdata)
1588         {
1589                 if ($curlResult->getHeader('server') == 'Mastodon') {
1590                         $serverdata['platform'] = 'mastodon';
1591                         $serverdata['network'] = Protocol::ACTIVITYPUB;
1592                 } elseif ($curlResult->inHeader('x-diaspora-version')) {
1593                         $serverdata['platform'] = 'diaspora';
1594                         $serverdata['network'] = Protocol::DIASPORA;
1595                         $serverdata['version'] = $curlResult->getHeader('x-diaspora-version');
1596                 } elseif ($curlResult->inHeader('x-friendica-version')) {
1597                         $serverdata['platform'] = 'friendica';
1598                         $serverdata['network'] = Protocol::DFRN;
1599                         $serverdata['version'] = $curlResult->getHeader('x-friendica-version');
1600                 } else {
1601                         return $serverdata;
1602                 }
1603
1604                 if ($serverdata['detection-method'] == self::DETECT_MANUAL) {
1605                         $serverdata['detection-method'] = self::DETECT_HEADER;
1606                 }
1607
1608                 return $serverdata;
1609         }
1610
1611         /**
1612          * Test if the body contains valid content
1613          *
1614          * @param string $body
1615          * @return boolean
1616          */
1617         private static function invalidBody(string $body)
1618         {
1619                 // Currently we only test for a HTML element.
1620                 // Possibly we enhance this in the future.
1621                 return !strpos($body, '>');
1622         }
1623
1624         /**
1625          * Update GServer entries
1626          */
1627         public static function discover()
1628         {
1629                 // Update the server list
1630                 self::discoverFederation();
1631
1632                 $no_of_queries = 5;
1633
1634                 $requery_days = intval(DI::config()->get('system', 'poco_requery_days'));
1635
1636                 if ($requery_days == 0) {
1637                         $requery_days = 7;
1638                 }
1639
1640                 $last_update = date('c', time() - (60 * 60 * 24 * $requery_days));
1641
1642                 $gservers = DBA::p("SELECT `id`, `url`, `nurl`, `network`, `poco`, `directory-type`
1643                         FROM `gserver`
1644                         WHERE NOT `failed`
1645                         AND `directory-type` != ?
1646                         AND `last_poco_query` < ?
1647                         ORDER BY RAND()", self::DT_NONE, $last_update
1648                 );
1649
1650                 while ($gserver = DBA::fetch($gservers)) {
1651                         Logger::info('Update peer list', ['server' => $gserver['url'], 'id' => $gserver['id']]);
1652                         Worker::add(PRIORITY_LOW, 'UpdateServerPeers', $gserver['url']);
1653
1654                         Logger::info('Update directory', ['server' => $gserver['url'], 'id' => $gserver['id']]);
1655                         Worker::add(PRIORITY_LOW, 'UpdateServerDirectory', $gserver);
1656
1657                         $fields = ['last_poco_query' => DateTimeFormat::utcNow()];
1658                         DBA::update('gserver', $fields, ['nurl' => $gserver['nurl']]);
1659
1660                         if (--$no_of_queries == 0) {
1661                                 break;
1662                         }
1663                 }
1664
1665                 DBA::close($gservers);
1666         }
1667
1668         /**
1669          * Discover federated servers
1670          */
1671         private static function discoverFederation()
1672         {
1673                 $last = DI::config()->get('poco', 'last_federation_discovery');
1674
1675                 if ($last) {
1676                         $next = $last + (24 * 60 * 60);
1677
1678                         if ($next > time()) {
1679                                 return;
1680                         }
1681                 }
1682
1683                 // Discover federated servers
1684                 $protocols = ['activitypub', 'diaspora', 'dfrn', 'ostatus'];
1685                 foreach ($protocols as $protocol) {
1686                         $query = '{nodes(protocol:"' . $protocol . '"){host}}';
1687                         $curlResult = DI::httpRequest()->fetch('https://the-federation.info/graphql?query=' . urlencode($query));
1688                         if (!empty($curlResult)) {
1689                                 $data = json_decode($curlResult, true);
1690                                 if (!empty($data['data']['nodes'])) {
1691                                         foreach ($data['data']['nodes'] as $server) {
1692                                                 // Using "only_nodeinfo" since servers that are listed on that page should always have it.
1693                                                 Worker::add(PRIORITY_LOW, 'UpdateGServer', 'https://' . $server['host'], true);
1694                                         }
1695                                 }
1696                         }
1697                 }
1698
1699                 // Disvover Mastodon servers
1700                 $accesstoken = DI::config()->get('system', 'instances_social_key');
1701
1702                 if (!empty($accesstoken)) {
1703                         $api = 'https://instances.social/api/1.0/instances/list?count=0';
1704                         $header = ['Authorization: Bearer '.$accesstoken];
1705                         $curlResult = DI::httpRequest()->get($api, ['header' => $header]);
1706
1707                         if ($curlResult->isSuccess()) {
1708                                 $servers = json_decode($curlResult->getBody(), true);
1709
1710                                 foreach ($servers['instances'] as $server) {
1711                                         $url = (is_null($server['https_score']) ? 'http' : 'https') . '://' . $server['name'];
1712                                         Worker::add(PRIORITY_LOW, 'UpdateGServer', $url);
1713                                 }
1714                         }
1715                 }
1716
1717                 DI::config()->set('poco', 'last_federation_discovery', time());
1718         }
1719 }