russian encoding #324

oscarotero · oscarotero · commit dc8125e5ec62 · 2019-07-18T13:33:13.000+02:00
diff --git a/src/Http/Response.php b/src/Http/Response.php
@@ -57,51 +57,7 @@ public function getHtmlContent()
                     return $this->htmlContent = false;
                 }
 
-                $errors = libxml_use_internal_errors(true);
-                $entities = libxml_disable_entity_loader(true);
-
-                $this->htmlContent = new DOMDocument();
-
-                if (stripos($content, '<meta charset="utf') === false) {
-                    $encodings = [
-                        'ASCII'        => 'ascii',
-                        'UTF-8'        => 'utf-8',
-                        'SJIS'         => 'shift_jis',
-                        'Windows-1251' => 'windows-1251',
-                        'Windows-1252' => 'windows-1252',
-                        'Windows-1254' => 'windows-1254',
-                        'ISO-8859-16'  => 'iso-8859-16',
-                        'ISO-8859-15'  => 'iso-8859-15',
-                        'ISO-8859-14'  => 'iso-8859-14',
-                        'ISO-8859-13'  => 'iso-8859-13',
-                        'ISO-8859-10'  => 'iso-8859-10',
-                        'ISO-8859-9'   => 'iso-8859-9',
-                        'ISO-8859-8'   => 'iso-8859-8',
-                        'ISO-8859-7'   => 'iso-8859-7',
-                        'ISO-8859-6'   => 'iso-8859-6',
-                        'ISO-8859-5'   => 'iso-8859-5',
-                        'ISO-8859-4'   => 'iso-8859-4',
-                        'ISO-8859-3'   => 'iso-8859-3',
-                        'ISO-8859-2'   => 'iso-8859-2',
-                        'ISO-8859-1'   => 'iso-8859-1', 
-                    ];
-                    
-                    $detected = mb_detect_encoding($content, implode(',', array_keys($encodings)), true);
-
-                    if ($detected && !empty($encodings[$detected])) {
-                        $content = mb_convert_encoding($content, 'HTML-ENTITIES', $detected);
-                        $content = preg_replace(
-                            '/<head[^>]*>/',
-                            '<head><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset='.$encodings[$detected].'">',
-                            $content
-                        );
-                    }
-                }
-
-                $this->htmlContent->loadHTML(trim($content));
-
-                libxml_use_internal_errors($errors);
-                libxml_disable_entity_loader($entities);
+                $this->htmlContent = Utils::parse($content);
             } catch (Exception $exception) {
                 return $this->htmlContent = false;
             }
diff --git a/src/Utils.php b/src/Utils.php
@@ -12,6 +12,29 @@
  */
 class Utils
 {
+    const ENCODINGS = [
+        'ASCII'        => 'ascii',
+        'UTF-8'        => 'utf-8',
+        'SJIS'         => 'shift_jis',
+        'Windows-1251' => 'windows-1251',
+        'Windows-1252' => 'windows-1252',
+        'Windows-1254' => 'windows-1254',
+        'ISO-8859-1'   => 'iso-8859-1',
+        'ISO-8859-2'   => 'iso-8859-2',
+        'ISO-8859-3'   => 'iso-8859-3',
+        'ISO-8859-4'   => 'iso-8859-4',
+        'ISO-8859-5'   => 'iso-8859-5',
+        'ISO-8859-6'   => 'iso-8859-6',
+        'ISO-8859-7'   => 'iso-8859-7',
+        'ISO-8859-8'   => 'iso-8859-8',
+        'ISO-8859-9'   => 'iso-8859-9',
+        'ISO-8859-10'  => 'iso-8859-10',
+        'ISO-8859-13'  => 'iso-8859-13',
+        'ISO-8859-14'  => 'iso-8859-14',
+        'ISO-8859-15'  => 'iso-8859-15',
+        'ISO-8859-16'  => 'iso-8859-16',
+    ];
+
     /**
      * Creates a <video> element.
      *
@@ -240,4 +263,54 @@ public static function xpathQuery(DOMDocument $document, $query, $returnFirst =
             return $returnFirst ? $entries->item(0) : $entries;
         }
     }
+
+    /**
+     * Parse a string as html code
+     *
+     * @param string $html
+     *
+     * @return DOMDocument
+     */
+    public static function parse($html)
+    {
+        $errors = libxml_use_internal_errors(true);
+        $entities = libxml_disable_entity_loader(true);
+
+        $html = trim(self::normalize($html));
+
+        $document = new DOMDocument();
+        $document->loadHTML($html);
+
+        libxml_use_internal_errors($errors);
+        libxml_disable_entity_loader($entities);
+
+        return $document;
+    }
+
+    /**
+     * Normalize the encoding of a html code before parse
+     * 
+     * @param string $string
+     * 
+     * @return string
+     */
+    private static function normalize($string)
+    {
+        if (stripos($string, '<meta charset="utf') === false) {
+            return $string;
+        }
+
+        $detected = mb_detect_encoding($string, implode(',', array_keys(self::ENCODINGS)), true);
+        
+        if ($detected && isset(self::ENCODINGS[$detected])) {
+            $string = mb_convert_encoding($string, 'HTML-ENTITIES', $detected);
+            $string = preg_replace(
+                '/<head[^>]*>/',
+                '<head><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset='.self::ENCODINGS[$detected].'">',
+                $string
+            );
+        }
+
+        return $string;
+    }
 }
