2.0 migration note for duk_char_code_at()

doc/release-notes-v2-0.rst

@@ -985,6 +985,11 @@ Other incompatible changes
 * If a user function is called using the identifier 'eval', such a call won't
   get tailcall optimized even if otherwise possible.
 
+* ``duk_char_code_at()`` and ``String.charCodeAt()`` now return 0xFFFD (Unicode
+  replacement character) if the string cannot be decoded as extended UTF-8,
+  previously an error was thrown.  This situation never occurs for standard
+  Ecmascript strings or valid UTF-8 strings.
+
 Known issues
 ============