There is some related discussion on this in
this thread. Some relevant take-aways are that even in the context of the most common standard formatting, it is hard™️ to get correct and not trivial to optimize (see:
the great amount of effort put into to getting what we have today right). I'm not sure that's how we want the core team to be spending their time compared to other language features, so I definitely prefer keeping the maintenance burden slim here and outsourcing more robust, configurable, quicker-to-iterate parsers in libraries.