Best way to parse ascii letters and numbers

[DamianReeves]

I am implementing a parser for the package url (purl) spec and in looking at the type section, I need to deal with ASCII letters and numbers. What is the best way to construct the parser for type?

type:

The package type is composed only of ASCII letters and numbers, '.', '+' and '-' (period, plus, and dash)
The type cannot start with a number
The type cannot contains spaces
The type must NOT be percent-encoded
The type is case insensitive. The canonical form is lowercase

[j-mie6]

It's not clear what exactly it means by "not be percent encoded"; whether that's supposed to be "no percent-encoded characters in the name" or what. But basically, it seems like a direct translation is probably best:

val start = oneOf('a' to 'z') | oneOf('A' to 'Z') | oneOf('.', '+', '-')
val letters = oneOf('0' to '9') | start

(you could use digit/letter, but that would be more than just ASCII in this case; depends on how strict you want to be. Or you could build a set of all the letters and use a single oneOf, which would be more efficient in this case). Then you want to stitch them together, which can probably be done with span for simplicity:

val typePurl = (start ~> many(letters)).span.map(_.toLowerCase)

That will give you a canonicalisation to lowercase strings (I might have gotten the specific method wrong from Char, but you get the idea.

Then the percent-encoded bit should be trivially satisfied by not allowing for any % in the name... if I understand it correctly? Otherwise, you could finish it off with a filter on the typePurl to check for whatever is needed inside?

[DamianReeves]

Thanks so much for the quick response. This is awesome and makes sense.