X-Git-Url: https://git.mxchange.org/?a=blobdiff_plain;f=src%2FModel%2FGServer.php;h=c0e4899923d86cb375109e9b2240675d166cc5eb;hb=a662245c744f4d2faa1b533977f7d21b4de6a724;hp=3a2a0bd798b8115b9e34aeed0290c93feb9a9539;hpb=4719af67245293c94bd4c4cb75e242841ed32ab6;p=friendica.git diff --git a/src/Model/GServer.php b/src/Model/GServer.php index 3a2a0bd798..c0e4899923 100644 --- a/src/Model/GServer.php +++ b/src/Model/GServer.php @@ -1,6 +1,6 @@ '\_', '%' => '\%', '?' => '_', '*' => '%']); + + // The SUBSTRING_INDEX returns everything before the eventual third /, which effectively trims an + // eventual server path and keep only the server domain which we're matching against the pattern. + $sql = "SELECT `gserver`.*, COUNT(*) AS `contacts` + FROM `gserver` + LEFT JOIN `contact` ON `gserver`.`id` = `contact`.`gsid` + WHERE SUBSTRING_INDEX(`gserver`.`nurl`, '/', 3) LIKE ? + AND NOT `gserver`.`failed` + GROUP BY `gserver`.`id`"; + + $stmt = DI::dba()->p($sql, $likePattern); + + return DI::dba()->toArray($stmt); + } + /** * Checks if the given server is reachable * @@ -123,59 +172,53 @@ class GServer return self::check($server, $network, $force); } - /** - * Decides if a server needs to be updated, based upon several date fields - * - * @param date $created Creation date of that server entry - * @param date $updated When had the server entry be updated - * @param date $last_failure Last failure when contacting that server - * @param date $last_contact Last time the server had been contacted - * - * @return boolean Does the server record needs an update? - */ - public static function updateNeeded($created, $updated, $last_failure, $last_contact) + public static function getNextUpdateDate(bool $success, string $created = '', string $last_contact = '') { + // On successful contact process check again next week + if ($success) { + return DateTimeFormat::utc('now +7 day'); + } + $now = strtotime(DateTimeFormat::utcNow()); - if ($updated > $last_contact) { - $contact_time = strtotime($updated); + if ($created > $last_contact) { + $contact_time = strtotime($created); } else { $contact_time = strtotime($last_contact); } - $failure_time = strtotime($last_failure); - $created_time = strtotime($created); + // If the last contact was less than 6 hours before then try again in 6 hours + if (($now - $contact_time) < (60 * 60 * 6)) { + return DateTimeFormat::utc('now +6 hour'); + } - // If there is no "created" time then use the current time - if ($created_time <= 0) { - $created_time = $now; + // If the last contact was less than 12 hours before then try again in 12 hours + if (($now - $contact_time) < (60 * 60 * 12)) { + return DateTimeFormat::utc('now +12 hour'); } - // If the last contact was less than 24 hours then don't update + // If the last contact was less than 24 hours before then try tomorrow again if (($now - $contact_time) < (60 * 60 * 24)) { - return false; + return DateTimeFormat::utc('now +1 day'); } - // If the last failure was less than 24 hours then don't update - if (($now - $failure_time) < (60 * 60 * 24)) { - return false; + // If the last contact was less than a week before then try again in a week + if (($now - $contact_time) < (60 * 60 * 24 * 7)) { + return DateTimeFormat::utc('now +1 week'); } - // If the last contact was less than a week ago and the last failure is older than a week then don't update - //if ((($now - $contact_time) < (60 * 60 * 24 * 7)) && ($contact_time > $failure_time)) - // return false; - - // If the last contact time was more than a week ago and the contact was created more than a week ago, then only try once a week - if ((($now - $contact_time) > (60 * 60 * 24 * 7)) && (($now - $created_time) > (60 * 60 * 24 * 7)) && (($now - $failure_time) < (60 * 60 * 24 * 7))) { - return false; + // If the last contact was less than two weeks before then try again in two week + if (($now - $contact_time) < (60 * 60 * 24 * 14)) { + return DateTimeFormat::utc('now +2 week'); } - // If the last contact time was more than a month ago and the contact was created more than a month ago, then only try once a month - if ((($now - $contact_time) > (60 * 60 * 24 * 30)) && (($now - $created_time) > (60 * 60 * 24 * 30)) && (($now - $failure_time) < (60 * 60 * 24 * 30))) { - return false; + // If the last contact was less than a month before then try again in a month + if (($now - $contact_time) < (60 * 60 * 24 * 30)) { + return DateTimeFormat::utc('now +1 month'); } - return true; + // The system hadn't been successul contacted for more than a month, so try again in three months + return DateTimeFormat::utc('now +3 month'); } /** @@ -191,7 +234,6 @@ class GServer public static function check(string $server_url, string $network = '', bool $force = false, bool $only_nodeinfo = false) { $server_url = self::cleanURL($server_url); - if ($server_url == '') { return false; } @@ -204,24 +246,11 @@ class GServer DBA::update('gserver', $fields, $condition); } - $last_contact = $gserver['last_contact']; - $last_failure = $gserver['last_failure']; - - // See discussion under https://forum.friendi.ca/display/0b6b25a8135aabc37a5a0f5684081633 - // It can happen that a zero date is in the database, but storing it again is forbidden. - if ($last_contact < DBA::NULL_DATETIME) { - $last_contact = DBA::NULL_DATETIME; - } - - if ($last_failure < DBA::NULL_DATETIME) { - $last_failure = DBA::NULL_DATETIME; - } - - if (!$force && !self::updateNeeded($gserver['created'], '', $last_failure, $last_contact)) { + if (!$force && (strtotime($gserver['next_contact']) > time())) { Logger::info('No update needed', ['server' => $server_url]); - return ($last_contact >= $last_failure); + return (!$gserver['failed']); } - Logger::info('Server is outdated. Start discovery.', ['Server' => $server_url, 'Force' => $force, 'Created' => $gserver['created'], 'Failure' => $last_failure, 'Contact' => $last_contact]); + Logger::info('Server is outdated. Start discovery.', ['Server' => $server_url, 'Force' => $force]); } else { Logger::info('Server is unknown. Start discovery.', ['Server' => $server_url]); } @@ -234,10 +263,13 @@ class GServer * * @param string $url */ - private static function setFailure(string $url) + public static function setFailure(string $url) { - if (DBA::exists('gserver', ['nurl' => Strings::normaliseLink($url)])) { - DBA::update('gserver', ['failed' => true, 'last_failure' => DateTimeFormat::utcNow(), 'detection-method' => null], + $gserver = DBA::selectFirst('gserver', [], ['nurl' => Strings::normaliseLink($url)]); + if (DBA::isResult($gserver)) { + $next_update = self::getNextUpdateDate(false, $gserver['created'], $gserver['last_contact']); + DBA::update('gserver', ['failed' => true, 'last_failure' => DateTimeFormat::utcNow(), + 'next_contact' => $next_update, 'detection-method' => null], ['nurl' => Strings::normaliseLink($url)]); Logger::info('Set failed status for existing server', ['url' => $url]); return; @@ -305,18 +337,30 @@ class GServer // If the URL missmatches, then we mark the old entry as failure if ($url != $original_url) { + /// @todo What to do with "next_contact" here? DBA::update('gserver', ['failed' => true, 'last_failure' => DateTimeFormat::utcNow()], ['nurl' => Strings::normaliseLink($original_url)]); } // When a nodeinfo is present, we don't need to dig further $xrd_timeout = DI::config()->get('system', 'xrd_timeout'); - $curlResult = DI::httpRequest()->get($url . '/.well-known/nodeinfo', ['timeout' => $xrd_timeout]); + $curlResult = DI::httpClient()->get($url . '/.well-known/nodeinfo', HttpClientAccept::JSON, [HttpClientOptions::TIMEOUT => $xrd_timeout]); if ($curlResult->isTimeout()) { self::setFailure($url); return false; } + // On a redirect follow the new host but mark the old one as failure + if ($curlResult->isSuccess() && !empty($curlResult->getRedirectUrl()) && (parse_url($url, PHP_URL_HOST) != parse_url($curlResult->getRedirectUrl(), PHP_URL_HOST))) { + $curlResult = DI::httpClient()->get($url, HttpClientAccept::HTML, [HttpClientOptions::TIMEOUT => $xrd_timeout]); + if (!empty($curlResult->getRedirectUrl()) && parse_url($url, PHP_URL_HOST) != parse_url($curlResult->getRedirectUrl(), PHP_URL_HOST)) { + Logger::info('Found redirect. Mark old entry as failure', ['old' => $url, 'new' => $curlResult->getRedirectUrl()]); + self::setFailure($url); + self::detect($curlResult->getRedirectUrl(), $network, $only_nodeinfo); + return false; + } + } + $nodeinfo = self::fetchNodeinfo($url, $curlResult); if ($only_nodeinfo && empty($nodeinfo)) { Logger::info('Invalid nodeinfo in nodeinfo-mode, server is marked as failure', ['url' => $url]); @@ -334,6 +378,12 @@ class GServer if (empty($nodeinfo['network']) || in_array($nodeinfo['network'], [Protocol::DFRN, Protocol::ZOT])) { if (!empty($nodeinfo['detection-method'])) { $serverdata['detection-method'] = $nodeinfo['detection-method']; + + foreach (['registered-users', 'active_users_monthly', 'active-halfyear-users', 'local-posts'] as $field) { + if (!empty($nodeinfo[$field])) { + $serverdata[$field] = $nodeinfo[$field]; + } + } } // Fetch the landing page, possibly it reveals some data @@ -344,8 +394,15 @@ class GServer $basedata = ['detection-method' => self::DETECT_MANUAL]; } - $curlResult = DI::httpRequest()->get($baseurl, ['timeout' => $xrd_timeout]); + $curlResult = DI::httpClient()->get($baseurl, HttpClientAccept::HTML, [HttpClientOptions::TIMEOUT => $xrd_timeout]); if ($curlResult->isSuccess()) { + if (!empty($curlResult->getRedirectUrl()) && (parse_url($baseurl, PHP_URL_HOST) != parse_url($curlResult->getRedirectUrl(), PHP_URL_HOST))) { + Logger::info('Found redirect. Mark old entry as failure', ['old' => $url, 'new' => $curlResult->getRedirectUrl()]); + self::setFailure($url); + self::detect($curlResult->getRedirectUrl(), $network, $only_nodeinfo); + return false; + } + $basedata = self::analyseRootHeader($curlResult, $basedata); $basedata = self::analyseRootBody($curlResult, $basedata, $baseurl); } @@ -361,7 +418,7 @@ class GServer // When the base path doesn't seem to contain a social network we try the complete path. // Most detectable system have to be installed in the root directory. // We checked the base to avoid false positives. - $curlResult = DI::httpRequest()->get($url, ['timeout' => $xrd_timeout]); + $curlResult = DI::httpClient()->get($url, HttpClientAccept::HTML, [HttpClientOptions::TIMEOUT => $xrd_timeout]); if ($curlResult->isSuccess()) { $urldata = self::analyseRootHeader($curlResult, $serverdata); $urldata = self::analyseRootBody($curlResult, $urldata, $url); @@ -381,14 +438,20 @@ class GServer } } - if (empty($serverdata['network']) || ($serverdata['network'] == Protocol::ACTIVITYPUB)) { + if (empty($nodeinfo['network']) && (empty($serverdata['network']) || ($serverdata['network'] == Protocol::ACTIVITYPUB))) { $serverdata = self::detectMastodonAlikes($url, $serverdata); } // All following checks are done for systems that always have got a "host-meta" endpoint. // With this check we don't have to waste time and ressources for dead systems. // Also this hopefully prevents us from receiving abuse messages. - if (empty($serverdata['network']) && !self::validHostMeta($url)) { + $validHostMeta = self::validHostMeta($url); + + if (empty($serverdata['network']) && !$validHostMeta) { + $serverdata = self::detectFromContacts($url, $serverdata); + } + + if (empty($serverdata['network']) && !$validHostMeta) { self::setFailure($url); return false; } @@ -407,14 +470,22 @@ class GServer $serverdata = self::detectHubzilla($url, $serverdata); } + if (empty($serverdata['network']) || in_array($serverdata['detection-method'], [self::DETECT_MANUAL, self::DETECT_BODY])) { + $serverdata = self::detectPeertube($url, $serverdata); + } + if (empty($serverdata['network'])) { $serverdata = self::detectNextcloud($url, $serverdata); } - if (empty($serverdata['network'])) { + if (empty($nodeinfo['network']) && empty($serverdata['network'])) { $serverdata = self::detectGNUSocial($url, $serverdata); } + if (empty($serverdata['network'])) { + $serverdata = self::detectPumpIO($url, $serverdata); + } + $serverdata = array_merge($nodeinfo, $serverdata); } else { $serverdata = $nodeinfo; @@ -422,9 +493,13 @@ class GServer // Detect the directory type $serverdata['directory-type'] = self::DT_NONE; - $serverdata = self::checkPoCo($url, $serverdata); + $serverdata = self::checkMastodonDirectory($url, $serverdata); + if ($serverdata['directory-type'] == self::DT_NONE) { + $serverdata = self::checkPoCo($url, $serverdata); + } + // We can't detect the network type. Possibly it is some system that we don't know yet if (empty($serverdata['network'])) { $serverdata['network'] = Protocol::PHANTOM; @@ -438,19 +513,25 @@ class GServer $serverdata['url'] = $url; $serverdata['nurl'] = Strings::normaliseLink($url); - // We take the highest number that we do find - $registeredUsers = $serverdata['registered-users'] ?? 0; + if (in_array($serverdata['network'], [Protocol::PHANTOM, Protocol::FEED])) { + $serverdata = self::detectNetworkViaContacts($url, $serverdata); + } - // On an active server there has to be at least a single user - if (($serverdata['network'] != Protocol::PHANTOM) && ($registeredUsers == 0)) { - $registeredUsers = 1; + if ($serverdata['network'] == Protocol::ACTIVITYPUB) { + $serverdata = self::fetchWeeklyUsage($url, $serverdata); } - if ($serverdata['network'] == Protocol::PHANTOM) { - $serverdata['registered-users'] = $registeredUsers; - $serverdata = self::detectNetworkViaContacts($url, $serverdata); + $serverdata['registered-users'] = $serverdata['registered-users'] ?? 0; + + // On an active server there has to be at least a single user + if (!in_array($serverdata['network'], [Protocol::PHANTOM, Protocol::FEED]) && ($serverdata['registered-users'] == 0)) { + $serverdata['registered-users'] = 1; + } elseif (in_array($serverdata['network'], [Protocol::PHANTOM, Protocol::FEED])) { + $serverdata['registered-users'] = 0; } + $serverdata['next_contact'] = self::getNextUpdateDate(true); + $serverdata['last_contact'] = DateTimeFormat::utcNow(); $serverdata['failed'] = false; @@ -460,11 +541,6 @@ class GServer $ret = DBA::insert('gserver', $serverdata); $id = DBA::lastInsertId(); } else { - // Don't override the network with 'unknown' when there had been a valid entry before - if (($serverdata['network'] == Protocol::PHANTOM) && !empty($gserver['network'])) { - unset($serverdata['network']); - } - $ret = DBA::update('gserver', $serverdata, ['nurl' => $serverdata['nurl']]); $gserver = DBA::selectFirst('gserver', ['id'], ['nurl' => $serverdata['nurl']]); if (DBA::isResult($gserver)) { @@ -472,14 +548,31 @@ class GServer } } - if (!empty($serverdata['network']) && !empty($id) && ($serverdata['network'] != Protocol::PHANTOM)) { + // Count the number of known contacts from this server + if (!empty($id) && !in_array($serverdata['network'], [Protocol::PHANTOM, Protocol::FEED])) { $apcontacts = DBA::count('apcontact', ['gsid' => $id]); - $contacts = DBA::count('contact', ['uid' => 0, 'gsid' => $id]); - $max_users = max($apcontacts, $contacts, $registeredUsers); - if ($max_users > $registeredUsers) { + $contacts = DBA::count('contact', ['uid' => 0, 'gsid' => $id, 'failed' => false]); + $max_users = max($apcontacts, $contacts); + if ($max_users > $serverdata['registered-users']) { Logger::info('Update registered users', ['id' => $id, 'url' => $serverdata['nurl'], 'registered-users' => $max_users]); DBA::update('gserver', ['registered-users' => $max_users], ['id' => $id]); } + + if (empty($serverdata['active-month-users'])) { + $contacts = DBA::count('contact', ["`uid` = ? AND `gsid` = ? AND NOT `failed` AND `last-item` > ?", 0, $id, DateTimeFormat::utc('now - 30 days')]); + if ($contacts > 0) { + Logger::info('Update monthly users', ['id' => $id, 'url' => $serverdata['nurl'], 'monthly-users' => $contacts]); + DBA::update('gserver', ['active-month-users' => $contacts], ['id' => $id]); + } + } + + if (empty($serverdata['active-halfyear-users'])) { + $contacts = DBA::count('contact', ["`uid` = ? AND `gsid` = ? AND NOT `failed` AND `last-item` > ?", 0, $id, DateTimeFormat::utc('now - 180 days')]); + if ($contacts > 0) { + Logger::info('Update halfyear users', ['id' => $id, 'url' => $serverdata['nurl'], 'halfyear-users' => $contacts]); + DBA::update('gserver', ['active-halfyear-users' => $contacts], ['id' => $id]); + } + } } if (!empty($serverdata['network']) && in_array($serverdata['network'], [Protocol::DFRN, Protocol::DIASPORA])) { @@ -499,7 +592,7 @@ class GServer { Logger::info('Discover relay data', ['server' => $server_url]); - $curlResult = DI::httpRequest()->get($server_url . '/.well-known/x-social-relay'); + $curlResult = DI::httpClient()->get($server_url . '/.well-known/x-social-relay', HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return; } @@ -518,7 +611,7 @@ class GServer $data['tags'] = []; } - $gserver = DBA::selectFirst('gserver', ['id', 'relay-subscribe', 'relay-scope'], ['nurl' => Strings::normaliseLink($server_url)]); + $gserver = DBA::selectFirst('gserver', ['id', 'url', 'network', 'relay-subscribe', 'relay-scope'], ['nurl' => Strings::normaliseLink($server_url)]); if (!DBA::isResult($gserver)) { return; } @@ -541,7 +634,7 @@ class GServer } foreach ($tags as $tag) { - DBA::insert('gserver-tag', ['gserver-id' => $gserver['id'], 'tag' => $tag], true); + DBA::insert('gserver-tag', ['gserver-id' => $gserver['id'], 'tag' => $tag], Database::INSERT_IGNORE); } } @@ -567,8 +660,22 @@ class GServer $fields['batch'] = $data['protocols']['dfrn']; } } + + if (isset($data['protocols']['activitypub'])) { + $fields['network'] = Protocol::ACTIVITYPUB; + + if (!empty($data['protocols']['activitypub']['actor'])) { + $fields['url'] = $data['protocols']['activitypub']['actor']; + } + if (!empty($data['protocols']['activitypub']['receive'])) { + $fields['batch'] = $data['protocols']['activitypub']['receive']; + } + } } - Diaspora::setRelayContact($server_url, $fields); + + Logger::info('Discovery ended', ['server' => $server_url, 'data' => $fields]); + + Relay::updateContact($gserver, $fields); } /** @@ -580,7 +687,7 @@ class GServer */ private static function fetchStatistics(string $url) { - $curlResult = DI::httpRequest()->get($url . '/statistics.json'); + $curlResult = DI::httpClient()->get($url . '/statistics.json', HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return []; } @@ -617,6 +724,21 @@ class GServer } } + if (!empty($data['total_users'])) { + $serverdata['registered-users'] = max($data['total_users'], 1); + } + + if (!empty($data['active_users_monthly'])) { + $serverdata['active-month-users'] = max($data['active_users_monthly'], 0); + } + + if (!empty($data['active_users_halfyear'])) { + $serverdata['active-halfyear-users'] = max($data['active_users_halfyear'], 0); + } + + if (!empty($data['local_posts'])) { + $serverdata['local-posts'] = max($data['local_posts'], 0); + } if (!empty($data['registrations_open'])) { $serverdata['register_policy'] = Register::OPEN; @@ -630,18 +752,19 @@ class GServer /** * Detect server type by using the nodeinfo data * - * @param string $url address of the server - * @param CurlResult $curlResult + * @param string $url address of the server + * @param ICanHandleHttpResponses $httpResult + * * @return array Server data * @throws \Friendica\Network\HTTPException\InternalServerErrorException */ - private static function fetchNodeinfo(string $url, CurlResult $curlResult) + private static function fetchNodeinfo(string $url, ICanHandleHttpResponses $httpResult) { - if (!$curlResult->isSuccess()) { + if (!$httpResult->isSuccess()) { return []; } - $nodeinfo = json_decode($curlResult->getBody(), true); + $nodeinfo = json_decode($httpResult->getBody(), true); if (!is_array($nodeinfo) || empty($nodeinfo['links'])) { return []; @@ -690,8 +813,7 @@ class GServer */ private static function parseNodeinfo1(string $nodeinfo_url) { - $curlResult = DI::httpRequest()->get($nodeinfo_url); - + $curlResult = DI::httpClient()->get($nodeinfo_url, HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return []; } @@ -727,7 +849,23 @@ class GServer } if (!empty($nodeinfo['usage']['users']['total'])) { - $server['registered-users'] = $nodeinfo['usage']['users']['total']; + $server['registered-users'] = max($nodeinfo['usage']['users']['total'], 1); + } + + if (!empty($nodeinfo['usage']['users']['activeMonth'])) { + $server['active-month-users'] = max($nodeinfo['usage']['users']['activeMonth'], 0); + } + + if (!empty($nodeinfo['usage']['users']['activeHalfyear'])) { + $server['active-halfyear-users'] = max($nodeinfo['usage']['users']['activeHalfyear'], 0); + } + + if (!empty($nodeinfo['usage']['localPosts'])) { + $server['local-posts'] = max($nodeinfo['usage']['localPosts'], 0); + } + + if (!empty($nodeinfo['usage']['localComments'])) { + $server['local-comments'] = max($nodeinfo['usage']['localComments'], 0); } if (!empty($nodeinfo['protocols']['inbound']) && is_array($nodeinfo['protocols']['inbound'])) { @@ -761,13 +899,14 @@ class GServer /** * Parses Nodeinfo 2 * + * @see https://git.feneas.org/jaywink/nodeinfo2 * @param string $nodeinfo_url address of the nodeinfo path * @return array Server data * @throws \Friendica\Network\HTTPException\InternalServerErrorException */ private static function parseNodeinfo2(string $nodeinfo_url) { - $curlResult = DI::httpRequest()->get($nodeinfo_url); + $curlResult = DI::httpClient()->get($nodeinfo_url, HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return []; } @@ -795,6 +934,11 @@ class GServer // Version numbers on Nodeinfo are presented with additional info, e.g.: // 0.6.3.0-p1702cc1c, 0.6.99.0-p1b9ab160 or 3.4.3-2-1191. $server['version'] = preg_replace('=(.+)-(.{4,})=ism', '$1', $server['version']); + + // qoto advertises itself as Mastodon + if (($server['platform'] == 'mastodon') && substr($nodeinfo['software']['version'], -5) == '-qoto') { + $server['platform'] = 'qoto'; + } } } @@ -803,13 +947,31 @@ class GServer } if (!empty($nodeinfo['usage']['users']['total'])) { - $server['registered-users'] = $nodeinfo['usage']['users']['total']; + $server['registered-users'] = max($nodeinfo['usage']['users']['total'], 1); + } + + if (!empty($nodeinfo['usage']['users']['activeMonth'])) { + $server['active-month-users'] = max($nodeinfo['usage']['users']['activeMonth'], 0); + } + + if (!empty($nodeinfo['usage']['users']['activeHalfyear'])) { + $server['active-halfyear-users'] = max($nodeinfo['usage']['users']['activeHalfyear'], 0); + } + + if (!empty($nodeinfo['usage']['localPosts'])) { + $server['local-posts'] = max($nodeinfo['usage']['localPosts'], 0); + } + + if (!empty($nodeinfo['usage']['localComments'])) { + $server['local-comments'] = max($nodeinfo['usage']['localComments'], 0); } if (!empty($nodeinfo['protocols'])) { $protocols = []; foreach ($nodeinfo['protocols'] as $protocol) { - $protocols[$protocol] = true; + if (is_string($protocol)) { + $protocols[$protocol] = true; + } } if (!empty($protocols['dfrn'])) { @@ -844,7 +1006,7 @@ class GServer */ private static function fetchSiteinfo(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/siteinfo.json'); + $curlResult = DI::httpClient()->get($url . '/siteinfo.json', HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return $serverdata; } @@ -880,7 +1042,23 @@ class GServer } if (!empty($data['channels_total'])) { - $serverdata['registered-users'] = $data['channels_total']; + $serverdata['registered-users'] = max($data['channels_total'], 1); + } + + if (!empty($data['channels_active_monthly'])) { + $serverdata['active-month-users'] = max($data['channels_active_monthly'], 0); + } + + if (!empty($data['channels_active_halfyear'])) { + $serverdata['active-halfyear-users'] = max($data['channels_active_halfyear'], 0); + } + + if (!empty($data['local_posts'])) { + $serverdata['local-posts'] = max($data['local_posts'], 0); + } + + if (!empty($data['local_comments'])) { + $serverdata['local-comments'] = max($data['local_comments'], 0); } if (!empty($data['register_policy'])) { @@ -913,7 +1091,7 @@ class GServer private static function validHostMeta(string $url) { $xrd_timeout = DI::config()->get('system', 'xrd_timeout'); - $curlResult = DI::httpRequest()->get($url . '/.well-known/host-meta', ['timeout' => $xrd_timeout]); + $curlResult = DI::httpClient()->get($url . '/.well-known/host-meta', HttpClientAccept::XRD_XML, [HttpClientOptions::TIMEOUT => $xrd_timeout]); if (!$curlResult->isSuccess()) { return false; } @@ -976,16 +1154,18 @@ class GServer return $serverdata; } + $time = time(); foreach ($contacts as $contact) { - $probed = Contact::getByURL($contact); - if (!empty($probed) && in_array($probed['network'], Protocol::FEDERATED)) { + $probed = Contact::getByURL($contact, true); + if (!empty($probed) && !$probed['failed'] && in_array($probed['network'], Protocol::FEDERATED)) { $serverdata['network'] = $probed['network']; break; + } elseif ((time() - $time) > 10) { + // To reduce the stress on remote systems we probe a maximum of 10 seconds + break; } } - $serverdata['registered-users'] = max($serverdata['registered-users'], count($contacts)); - return $serverdata; } @@ -1003,7 +1183,7 @@ class GServer { $serverdata['poco'] = ''; - $curlResult = DI::httpRequest()->get($url . '/poco'); + $curlResult = DI::httpClient()->get($url . '/poco', HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return $serverdata; } @@ -1015,7 +1195,7 @@ class GServer if (!empty($data['totalResults'])) { $registeredUsers = $serverdata['registered-users'] ?? 0; - $serverdata['registered-users'] = max($data['totalResults'], $registeredUsers); + $serverdata['registered-users'] = max($data['totalResults'], $registeredUsers, 1); $serverdata['directory-type'] = self::DT_POCO; $serverdata['poco'] = $url . '/poco'; } @@ -1033,7 +1213,7 @@ class GServer */ public static function checkMastodonDirectory(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/api/v1/directory?limit=1'); + $curlResult = DI::httpClient()->get($url . '/api/v1/directory?limit=1', HttpClientAccept::JSON); if (!$curlResult->isSuccess()) { return $serverdata; } @@ -1050,6 +1230,53 @@ class GServer return $serverdata; } + /** + * Detects Peertube via their known endpoint + * + * @param string $url URL of the given server + * @param array $serverdata array with server data + * + * @return array server data + */ + private static function detectPeertube(string $url, array $serverdata) + { + $curlResult = DI::httpClient()->get($url . '/api/v1/config', HttpClientAccept::JSON); + if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) { + return $serverdata; + } + + $data = json_decode($curlResult->getBody(), true); + if (empty($data)) { + return $serverdata; + } + + if (!empty($data['instance']) && !empty($data['serverVersion'])) { + $serverdata['platform'] = 'peertube'; + $serverdata['version'] = $data['serverVersion']; + $serverdata['network'] = Protocol::ACTIVITYPUB; + + if (!empty($data['instance']['name'])) { + $serverdata['site_name'] = $data['instance']['name']; + } + + if (!empty($data['instance']['shortDescription'])) { + $serverdata['info'] = $data['instance']['shortDescription']; + } + + if (!empty($data['signup'])) { + if (!empty($data['signup']['allowed'])) { + $serverdata['register_policy'] = Register::OPEN; + } + } + + if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) { + $serverdata['detection-method'] = self::DETECT_V1_CONFIG; + } + } + + return $serverdata; + } + /** * Detects the version number of a given server when it was a NextCloud installation * @@ -1060,8 +1287,7 @@ class GServer */ private static function detectNextcloud(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/status.php'); - + $curlResult = DI::httpClient()->get($url . '/status.php', HttpClientAccept::JSON); if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) { return $serverdata; } @@ -1084,6 +1310,65 @@ class GServer return $serverdata; } + private static function fetchWeeklyUsage(string $url, array $serverdata) { + $curlResult = DI::httpClient()->get($url . '/api/v1/instance/activity', HttpClientAccept::JSON); + if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) { + return $serverdata; + } + + $data = json_decode($curlResult->getBody(), true); + if (empty($data)) { + return $serverdata; + } + + $current_week = []; + foreach ($data as $week) { + // Use only data from a full week + if (empty($week['week']) || (time() - $week['week']) < 7 * 24 * 60 * 60) { + continue; + } + + // Most likely the data is sorted correctly. But we better are safe than sorry + if (empty($current_week['week']) || ($current_week['week'] < $week['week'])) { + $current_week = $week; + } + } + + if (!empty($current_week['logins'])) { + $serverdata['active-week-users'] = max($current_week['logins'], 0); + } + + return $serverdata; + } + + /** + * Detects the server network type from contacts of that server + * + * @param string $url URL of the given server + * @param array $serverdata array with server data + * + * @return array server data + */ + private static function detectFromContacts(string $url, array $serverdata) + { + $gserver = DBA::selectFirst('gserver', ['id'], ['nurl' => Strings::normaliseLink($url)]); + if (empty($gserver)) { + return $serverdata; + } + + $contact = Contact::selectFirst(['id'], ['uid' => 0, 'failed' => false, 'gsid' => $gserver['id']]); + + // Via probing we can be sure that the server is responding + if (!empty($contact['id']) && Contact::updateFromProbe($contact['id'])) { + $contact = Contact::selectFirst(['network', 'failed'], ['id' => $contact['id']]); + if (!$contact['failed'] && in_array($contact['network'], Protocol::FEDERATED)) { + $serverdata['network'] = $contact['network']; + } + } + + return $serverdata; + } + /** * Detects data from a given server url if it was a mastodon alike system * @@ -1094,8 +1379,7 @@ class GServer */ private static function detectMastodonAlikes(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/api/v1/instance'); - + $curlResult = DI::httpClient()->get($url . '/api/v1/instance', HttpClientAccept::JSON); if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) { return $serverdata; } @@ -1129,7 +1413,7 @@ class GServer } if (!empty($data['stats']['user_count'])) { - $serverdata['registered-users'] = $data['stats']['user_count']; + $serverdata['registered-users'] = max($data['stats']['user_count'], 1); } if (!empty($serverdata['version']) && preg_match('/.*?\(compatible;\s(.*)\s(.*)\)/ism', $serverdata['version'], $matches)) { @@ -1160,7 +1444,7 @@ class GServer */ private static function detectHubzilla(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/api/statusnet/config.json'); + $curlResult = DI::httpClient()->get($url . '/api/statusnet/config.json', HttpClientAccept::JSON); if (!$curlResult->isSuccess() || ($curlResult->getBody() == '')) { return $serverdata; } @@ -1245,7 +1529,57 @@ class GServer } return $val; - } + } + + /** + * Detect if the URL belongs to a pump.io server + * + * @param string $url URL of the given server + * @param array $serverdata array with server data + * + * @return array server data + */ + private static function detectPumpIO(string $url, array $serverdata) + { + $curlResult = DI::httpClient()->get($url . '/.well-known/host-meta.json', HttpClientAccept::JSON); + if (!$curlResult->isSuccess()) { + return $serverdata; + } + + $data = json_decode($curlResult->getBody(), true); + if (empty($data['links'])) { + return $serverdata; + + } + + // We are looking for some endpoints that are typical for pump.io + $trust = 0; + foreach ($data['links'] as $link) { + if (empty($link['rel'])) { + continue; + } + if (in_array($link['rel'], ['registration_endpoint', 'dialback', 'http://apinamespace.org/activitypub/whoami'])) { + ++$trust; + } + } + + if ($trust == 3) { + $serverdata['detection-method'] = self::DETECT_PUMPIO; + + $serverdata['platform'] = 'pumpio'; + $serverdata['version'] = ''; + $serverdata['network'] = Protocol::PUMPIO; + + $servers = $curlResult->getHeader('Server'); + foreach ($servers as $server) { + if (preg_match("#pump.io/(.*)\s#U", $server, $matches)) { + $serverdata['version'] = $matches[1]; + } + } + } + + return $serverdata; + } /** * Detect if the URL belongs to a GNU Social server @@ -1258,7 +1592,7 @@ class GServer private static function detectGNUSocial(string $url, array $serverdata) { // Test for GNU Social - $curlResult = DI::httpRequest()->get($url . '/api/gnusocial/version.json'); + $curlResult = DI::httpClient()->get($url . '/api/gnusocial/version.json', HttpClientAccept::JSON); if ($curlResult->isSuccess() && ($curlResult->getBody() != '{"error":"not implemented"}') && ($curlResult->getBody() != '') && (strlen($curlResult->getBody()) < 30)) { $serverdata['platform'] = 'gnusocial'; @@ -1271,12 +1605,12 @@ class GServer if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) { $serverdata['detection-method'] = self::DETECT_GNUSOCIAL; } - + return $serverdata; } // Test for Statusnet - $curlResult = DI::httpRequest()->get($url . '/api/statusnet/version.json'); + $curlResult = DI::httpClient()->get($url . '/api/statusnet/version.json', HttpClientAccept::JSON); if ($curlResult->isSuccess() && ($curlResult->getBody() != '{"error":"not implemented"}') && ($curlResult->getBody() != '') && (strlen($curlResult->getBody()) < 30)) { @@ -1312,9 +1646,11 @@ class GServer */ private static function detectFriendica(string $url, array $serverdata) { - $curlResult = DI::httpRequest()->get($url . '/friendica/json'); + // There is a bug in some versions of Friendica that will return an ActivityStream actor when the content type "application/json" is requested. + // Because of this me must not use ACCEPT_JSON here. + $curlResult = DI::httpClient()->get($url . '/friendica/json'); if (!$curlResult->isSuccess()) { - $curlResult = DI::httpRequest()->get($url . '/friendika/json'); + $curlResult = DI::httpClient()->get($url . '/friendika/json'); $friendika = true; $platform = 'Friendika'; } else { @@ -1331,7 +1667,7 @@ class GServer return $serverdata; } - if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) { + if (in_array($serverdata['detection-method'], [self::DETECT_HEADER, self::DETECT_BODY, self::DETECT_MANUAL])) { $serverdata['detection-method'] = $friendika ? self::DETECT_FRIENDIKA : self::DETECT_FRIENDICA; } @@ -1386,6 +1722,14 @@ class GServer */ private static function analyseRootBody($curlResult, array $serverdata, string $url) { + if (empty($curlResult->getBody())) { + return $serverdata; + } + + // Using only body information we cannot safely detect a lot of systems. + // So we define a list of platforms that we can detect safely. + $valid_platforms = ['friendica', 'friendika', 'diaspora', 'mastodon', 'hubzilla', 'misskey', 'peertube', 'wordpress', 'write.as']; + $doc = new DOMDocument(); @$doc->loadHTML($curlResult->getBody()); $xpath = new DOMXPath($doc); @@ -1431,15 +1775,12 @@ class GServer if (count($version_part) == 2) { if (in_array($version_part[0], ['WordPress'])) { - $serverdata['platform'] = strtolower($version_part[0]); + $serverdata['platform'] = 'wordpress'; $serverdata['version'] = $version_part[1]; // We still do need a reliable test if some AP plugin is activated - if (DBA::exists('apcontact', ['baseurl' => $url])) { - $serverdata['network'] = Protocol::ACTIVITYPUB; - } else { - $serverdata['network'] = Protocol::FEED; - } + // By now we just check in a later process for some known contacts + $serverdata['network'] = Protocol::FEED; if ($serverdata['detection-method'] == self::DETECT_MANUAL) { $serverdata['detection-method'] = self::DETECT_BODY; @@ -1499,7 +1840,11 @@ class GServer } } - if (!empty($serverdata['network']) && ($serverdata['detection-method'] == self::DETECT_MANUAL)) { + if (!empty($serverdata['platform']) && in_array($serverdata['detection-method'], [self::DETECT_MANUAL, self::DETECT_BODY]) && !in_array($serverdata['platform'], $valid_platforms)) { + $serverdata['network'] = Protocol::PHANTOM; + $serverdata['version'] = ''; + $serverdata['detection-method'] = self::DETECT_MANUAL; + } elseif (!empty($serverdata['network']) && ($serverdata['detection-method'] == self::DETECT_MANUAL)) { $serverdata['detection-method'] = self::DETECT_BODY; } @@ -1522,11 +1867,11 @@ class GServer } elseif ($curlResult->inHeader('x-diaspora-version')) { $serverdata['platform'] = 'diaspora'; $serverdata['network'] = Protocol::DIASPORA; - $serverdata['version'] = $curlResult->getHeader('x-diaspora-version'); + $serverdata['version'] = $curlResult->getHeader('x-diaspora-version')[0] ?? ''; } elseif ($curlResult->inHeader('x-friendica-version')) { $serverdata['platform'] = 'friendica'; $serverdata['network'] = Protocol::DFRN; - $serverdata['version'] = $curlResult->getHeader('x-friendica-version'); + $serverdata['version'] = $curlResult->getHeader('x-friendica-version')[0] ?? ''; } else { return $serverdata; } @@ -1569,22 +1914,11 @@ class GServer $last_update = date('c', time() - (60 * 60 * 24 * $requery_days)); - $gservers = DBA::p("SELECT `id`, `url`, `nurl`, `network`, `poco`, `directory-type` - FROM `gserver` - WHERE NOT `failed` - AND `directory-type` != ? - AND `last_poco_query` < ? - ORDER BY RAND()", self::DT_NONE, $last_update - ); + $gservers = DBA::select('gserver', ['id', 'url', 'nurl', 'network', 'poco', 'directory-type'], + ["NOT `failed` AND `directory-type` != ? AND `last_poco_query` < ?", GServer::DT_NONE, $last_update], + ['order' => ['RAND()']]); while ($gserver = DBA::fetch($gservers)) { - if (!GServer::check($gserver['url'], $gserver['network'])) { - // The server is not reachable? Okay, then we will try it later - $fields = ['last_poco_query' => DateTimeFormat::utcNow()]; - DBA::update('gserver', $fields, ['nurl' => $gserver['nurl']]); - continue; - } - Logger::info('Update peer list', ['server' => $gserver['url'], 'id' => $gserver['id']]); Worker::add(PRIORITY_LOW, 'UpdateServerPeers', $gserver['url']); @@ -1593,7 +1927,7 @@ class GServer $fields = ['last_poco_query' => DateTimeFormat::utcNow()]; DBA::update('gserver', $fields, ['nurl' => $gserver['nurl']]); - + if (--$no_of_queries == 0) { break; } @@ -1621,13 +1955,13 @@ class GServer $protocols = ['activitypub', 'diaspora', 'dfrn', 'ostatus']; foreach ($protocols as $protocol) { $query = '{nodes(protocol:"' . $protocol . '"){host}}'; - $curlResult = DI::httpRequest()->fetch('https://the-federation.info/graphql?query=' . urlencode($query)); + $curlResult = DI::httpClient()->fetch('https://the-federation.info/graphql?query=' . urlencode($query), HttpClientAccept::JSON); if (!empty($curlResult)) { $data = json_decode($curlResult, true); if (!empty($data['data']['nodes'])) { foreach ($data['data']['nodes'] as $server) { // Using "only_nodeinfo" since servers that are listed on that page should always have it. - Worker::add(PRIORITY_LOW, 'UpdateGServer', 'https://' . $server['host'], true); + self::add('https://' . $server['host'], true); } } } @@ -1638,19 +1972,99 @@ class GServer if (!empty($accesstoken)) { $api = 'https://instances.social/api/1.0/instances/list?count=0'; - $header = ['Authorization: Bearer '.$accesstoken]; - $curlResult = DI::httpRequest()->get($api, ['header' => $header]); - + $curlResult = DI::httpClient()->get($api, HttpClientAccept::JSON, [HttpClientOptions::HEADERS => ['Authorization' => ['Bearer ' . $accesstoken]]]); if ($curlResult->isSuccess()) { $servers = json_decode($curlResult->getBody(), true); foreach ($servers['instances'] as $server) { $url = (is_null($server['https_score']) ? 'http' : 'https') . '://' . $server['name']; - Worker::add(PRIORITY_LOW, 'UpdateGServer', $url); + self::add($url); } } } DI::config()->set('poco', 'last_federation_discovery', time()); } + + /** + * Set the protocol for the given server + * + * @param int $gsid Server id + * @param int $protocol Protocol id + * @return void + * @throws Exception + */ + public static function setProtocol(int $gsid, int $protocol) + { + if (empty($gsid)) { + return; + } + + $gserver = DBA::selectFirst('gserver', ['protocol', 'url'], ['id' => $gsid]); + if (!DBA::isResult($gserver)) { + return; + } + + $old = $gserver['protocol']; + + if (!is_null($old)) { + /* + The priority for the protocols is: + 1. ActivityPub + 2. DFRN via Diaspora + 3. Legacy DFRN + 4. Diaspora + 5. OStatus + */ + + // We don't need to change it when nothing is to be changed + if ($old == $protocol) { + return; + } + + // We don't want to mark a server as OStatus when it had been marked with any other protocol before + if ($protocol == Post\DeliveryData::OSTATUS) { + return; + } + + // If the server is marked as ActivityPub then we won't change it to anything different + if ($old == Post\DeliveryData::ACTIVITYPUB) { + return; + } + + // Don't change it to anything lower than DFRN if the new one wasn't ActivityPub + if (($old == Post\DeliveryData::DFRN) && ($protocol != Post\DeliveryData::ACTIVITYPUB)) { + return; + } + + // Don't change it to Diaspora when it is a legacy DFRN server + if (($old == Post\DeliveryData::LEGACY_DFRN) && ($protocol == Post\DeliveryData::DIASPORA)) { + return; + } + } + + Logger::info('Protocol for server', ['protocol' => $protocol, 'old' => $old, 'id' => $gsid, 'url' => $gserver['url'], 'callstack' => System::callstack(20)]); + DBA::update('gserver', ['protocol' => $protocol], ['id' => $gsid]); + } + + /** + * Fetch the protocol of the given server + * + * @param int $gsid Server id + * @return int + * @throws Exception + */ + public static function getProtocol(int $gsid) + { + if (empty($gsid)) { + return null; + } + + $gserver = DBA::selectFirst('gserver', ['protocol'], ['id' => $gsid]); + if (DBA::isResult($gserver)) { + return $gserver['protocol']; + } + + return null; + } }