The current shortcode parser is based on a single regular expression, making it error-prone and incapable of many useful features, such as nesting two instances of the same shortcode.
While I still think it’s a good idea, I haven’t seen anyone complain about running into any problems on CP with the current shortcode implementation. So I think this can safely be closed.
If the problem is reported at some time in the future, we could potentially revisit this then.