Isn't this already covered in the cursoricon tests?
I don't think so. That test would be marked TODO_WINE in that case, so your patch would cause a test failure(unexpected success). If that test wasn't marked todo the behavior your patch adds would be implemented already.