use Friendica\Content\OEmbed;
use Friendica\Core\Hook;
use Friendica\Core\Logger;
+use Friendica\Database\Database;
use Friendica\Database\DBA;
+use Friendica\DI;
/**
* Get information about a given URL
'oembed' => $do_oembed, 'content' => serialize($data),
'created' => DateTimeFormat::utcNow()
],
- true
+ Database::INSERT_UPDATE
);
return $data;
return $siteinfo;
}
- $curlResult = Network::curl($url);
+ $curlResult = DI::httpRequest()->get($url);
if (!$curlResult->isSuccess()) {
return $siteinfo;
}
}
}
- // Fetch the first mentioned charset. Can be in body or header
$charset = '';
- if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
+ // Look for a charset, first in headers
+ // Expected form: Content-Type: text/html; charset=ISO-8859-4
+ if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $header, $matches)) {
$charset = trim(trim(trim(array_pop($matches)), ';,'));
}
+ // Then in body that gets precedence
+ // Expected forms:
+ // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+ // - <meta charset="utf-8">
+ // - <meta charset=utf-8>
+ // - <meta charSet="utf-8">
+ // We escape <style> and <script> tags since they can contain irrelevant charset information
+ // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
+ Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
+ if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
+ $charset = trim(trim(trim(array_pop($matches)), ';,'));
+ }
+ });
+
+ $siteinfo['charset'] = $charset;
+
if ($charset && strtoupper($charset) != 'UTF-8') {
// See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
$charset = str_ireplace('latin-1', 'latin1', $charset);