Add test for double replacement with a reversed order since the order within the dictionary actually matters.
Remove upper-case umlaut replacements from standard test since the string replacement happens after the input text is converted to lower-case, so these string replacements are not used.
Add additional tests for umlaut replacements without a conversion to lower-case to ensure that the replacement of upper-case characters is independent of the lower-case characters.