Improve name encoding support for platform 3

Most name strings should be encoded with UTF-16BE per the spec, but there are situations where other encodings are required or acceptable. This change only addresses a subset of potential encodings. fixes #70
dompdf · Dec 12, 2023 · 84c64a7 · 84c64a7
1 parent c5f7810
commit 84c64a7
Show file tree

Hide file tree

Showing 2 changed files with 42 additions and 3 deletions.
diff --git a/src/FontLib/Table/Type/name.php b/src/FontLib/Table/Type/name.php
@@ -150,11 +150,49 @@ protected function _parse() {
       $records[] = $record;
     }
 
+    $system_encodings = mb_list_encodings();
+    foreach ($system_encodings as $enc) {
+        $system_encodings = array_merge($system_encodings, @mb_encoding_aliases($enc));
+    }
+    $system_encodings = array_change_key_case(array_fill_keys($system_encodings, true), CASE_UPPER);
+
     $names = array();
     foreach ($records as $record) {
       $font->seek($tableOffset + $data["stringOffset"] + $record->offset);
-      $s                      = $font->read($record->length);
-      $record->string         = Font::UTF16ToUTF8($s);
+      $record->stringRaw = $font->read($record->length);
+
+      $encoding = "UTF-16";
+      switch ($record->platformID) {
+        case 3:
+          switch ($record->platformSpecificID) {
+            case 2:
+              if (\array_key_exists("SJIS", $system_encodings)) {
+                $encoding = mb_detect_encoding($record->stringRaw, ["SJIS", "UTF-16"], true);
+              }
+              break;
+            case 3:
+              if (\array_key_exists("GB18030", $system_encodings)) {
+                $encoding = mb_detect_encoding($record->stringRaw, ["GB18030", "UTF-16"], true);
+              }
+              break;
+            case 4:
+              if (\array_key_exists("BIG-5", $system_encodings)) {
+                $encoding = mb_detect_encoding($record->stringRaw, ["BIG-5", "UTF-16"], true);
+              }
+              break;
+            case 5:
+              if (\array_key_exists("UHC", $system_encodings)) {
+                $encoding = mb_detect_encoding($record->stringRaw, ["UHC", "UTF-16"], true);
+              }
+              break;
+          }
+          break;
+      }
+      if ($encoding === false) {
+        $encoding = "UTF-16";
+      }
+
+      $record->string = mb_convert_encoding($record->stringRaw, "UTF-8", $encoding);
       $names[$record->nameID] = $record;
     }
 
@@ -184,7 +222,7 @@ protected function _encode() {
     }
 
     foreach ($records as $record) {
-      $str = $record->getUTF16();
+      $str = $record->stringRaw;
       $length += $font->write($str, mb_strlen($str, "8bit"));
     }
 

diff --git a/src/FontLib/Table/Type/nameRecord.php b/src/FontLib/Table/Type/nameRecord.php
@@ -23,6 +23,7 @@ class nameRecord extends BinaryStream {
   public $length;
   public $offset;
   public $string;
+  public $stringRaw;
 
   public static $format = array(
     "platformID"         => self::uint16,