maybe it just needs an open source project to make that sort of thing free.

That is definitely needed. Unfortunately, it's one of those problems like dates and encodings: a twisty maze of little exceptions all alike. That's not something that lends itself to the usual open source processes. As bad luck would have it, this is also a problem that people are less likely to come across than encodings, let alone dates. It is mostly businesses which are affected. It would be great to codify this knowledge Regexp::Common style — but it will require some very dedicated people to do a lot of thankless and mostly unnoticed work before the ball gets rolling.