Fix deprecation of mb_convert_encoding on HTML-ENTITIES

https://stackoverflow.com/questions/8218230/php-domdocument-loadhtml-not-encoding-utf-8-correctly
Victrid · May 15, 2024 · f8cf410 · f8cf410
1 parent d473c71
commit f8cf410
Showing 1 changed file with 9 additions and 2 deletions.
diff --git a/extension.php b/extension.php
@@ -97,14 +97,20 @@ public static function getCacheImageUri(string $url): string
         return FreshRSS_Context::userConf()->image_cache_url . $url;
     }
 
+    public static function small($string)
+{
+    return substr($string, 0, 20);
+}
+
     public static function cache_images(string $content): string
     {
         if (empty($content)) {
             return $content;
         }
         $doc = new DOMDocument();
         libxml_use_internal_errors(true); // prevent tag soup errors from showing
-        $doc->loadHTML(mb_convert_encoding($content, 'HTML-ENTITIES', 'UTF-8'));
+        $encoding = mb_detect_encoding($content);
+        $doc->loadHTML('<!DOCTYPE html><meta charset="'.$encoding.'">'.$content);
         $imgs = $doc->getElementsByTagName('img');
         foreach ($imgs as $img) {
             if ($img->hasAttribute('src')) {
@@ -129,7 +135,8 @@ public static function swapUris(string $content): string
         }
         $doc = new DOMDocument();
         libxml_use_internal_errors(true); // prevent tag soup errors from showing
-        $doc->loadHTML(mb_convert_encoding($content, 'HTML-ENTITIES', 'UTF-8'));
+        $encoding = mb_detect_encoding($content);
+        $doc->loadHTML('<!DOCTYPE html><meta charset="'.$encoding.'">'.$content);
         $imgs = $doc->getElementsByTagName('img');
         foreach ($imgs as $img) {
             if ($img->hasAttribute('src')) {