Regex literals tokenizes incorrectly

[PEZ]

Regular expression literals should tokenise as one form/s-expression, but they don't. E.g. #"foo" tokenizes as the the punc token # and a string "foo".

A way to see that this is happening is to type a regex literal as an argument to a function and watch how the signature help mini-hover, match the wrong active parameter:

[cfehse]

@PEZ

Something similar is happening with the @ character. The form:
(print @fehse)
will be parsed to this:

0: {type: "open", offset: 0, raw: "(", state: {…}}
1: {type: "id", offset: 1, raw: "print", state: {…}}
2: {type: "ws", offset: 6, raw: " ", state: {…}}
3: {type: "junk", offset: 7, raw: "@", state: {…}}
4: {type: "id", offset: 8, raw: "fehse", state: {…}}
5: {type: "close", offset: 14, raw: ")", state: {…}}