fix: handle mb_encoding_aliases() behavior differences between PHP 7.4 and 8.0+

Author: uzulla

src/Document.php

@@ -31,31 +31,56 @@ public function __construct(Extractor $extractor)
         preg_match('/charset=(?:"|\')?(.*?)(?=$|\s|;|"|\'|>)/i', $contentType, $match);
         if (!empty($match[1])) {
             $encoding = trim($match[1], ',');
-            try {
-                $ret = mb_encoding_aliases($encoding ?? '');
-                if ($ret === false) {
-                    $encoding = null;
-                }
-            } catch (\ValueError $exception) {
-                $encoding = null;
-            }
+            $encoding = $this->getValidEncoding($encoding);
         }
         if (is_null($encoding) && !empty($html)) {
             preg_match('/charset=(?:"|\')?(.*?)(?=$|\s|;|"|\'|>)/i', $html, $match);
             if (!empty($match[1])) {
                 $encoding = trim($match[1], ',');
+                $encoding = $this->getValidEncoding($encoding);
             }
+        }
+        $this->document = !empty($html) ? Parser::parse($html, $encoding) : new DOMDocument();
+        $this->initXPath();
+    }
+
+    /**
+     * Get valid encoding name if it exists, otherwise return null
+     *
+     * Uses mb_encoding_aliases() to verify the encoding is valid.
+     *
+     * TODO: When dropping PHP 7.4 support, remove the PHP_VERSION_ID < 80000 branch.
+     * PHP version differences:
+     * - PHP 7.4: mb_encoding_aliases() returns false for invalid encoding and throws Warning for empty string
+     * - PHP 8.0+: mb_encoding_aliases() throws ValueError for invalid/empty encoding
+     *
+     * @see https://www.php.net/manual/en/function.mb-encoding-aliases.php
+     */
+    private function getValidEncoding(?string $encoding): ?string
+    {
+        if (PHP_VERSION_ID < 80000) {
+            // PHP 7.4: Check return value (false = invalid encoding)
+            // Need to check empty() first to avoid Warning
+            // TODO: Remove this entire branch when PHP 7.4 support is dropped
+            if (empty($encoding)) {
+                return null;
+            }
+            $ret = mb_encoding_aliases($encoding);
+            if ($ret === false) {
+                return null;
+            } else {
+                return $encoding;
+            }
+        } else {
+            // PHP 8.0+: ValueError exception is thrown for invalid/empty encoding
             try {
-                $ret = mb_encoding_aliases($encoding ?? '');
-                if ($ret === false) {
-                    $encoding = null;
-                }
+                $aliases = mb_encoding_aliases($encoding ?? '');
+                // Check if aliases array is not empty (valid encoding should have at least one alias)
+                return !empty($aliases) ? $encoding : null;
             } catch (\ValueError $exception) {
-                $encoding = null;
+                return null;
             }
         }
-        $this->document = !empty($html) ? Parser::parse($html, $encoding) : new DOMDocument();
-        $this->initXPath();
     }
 
     private function initXPath()