Simplify + optimize lengthBytesUTF8. Don't fully decode surrogates (#…

…17413)
emscripten-core · Jul 17, 2022 · 8c047b1 · 8c047b1
1 parent 14c106a
commit 8c047b1
Showing 1 changed file with 10 additions and 6 deletions.
diff --git a/src/runtime_strings.js b/src/runtime_strings.js
@@ -191,12 +191,16 @@ function lengthBytesUTF8(str) {
   for (var i = 0; i < str.length; ++i) {
     // Gotcha: charCodeAt returns a 16-bit word that is a UTF-16 encoded code unit, not a Unicode code point of the character! So decode UTF16->UTF32->UTF8.
     // See http://unicode.org/faq/utf_bom.html#utf16-3
-    var u = str.charCodeAt(i); // possibly a lead surrogate
-    if (u >= 0xD800 && u <= 0xDFFF) u = 0x10000 + ((u & 0x3FF) << 10) | (str.charCodeAt(++i) & 0x3FF);
-    if (u <= 0x7F) ++len;
-    else if (u <= 0x7FF) len += 2;
-    else if (u <= 0xFFFF) len += 3;
-    else len += 4;
+    var c = str.charCodeAt(i); // possibly a lead surrogate
+    if (c <= 0x7F) {
+      len++;
+    } else if (c <= 0x7FF) {
+      len += 2;
+    } else if (c >= 0xD800 && c <= 0xDFFF) {
+      len += 4; ++i;
+    } else {
+      len += 3;
+    }
   }
   return len;
 }