Skip to content

Comment on Parsing – REBOL Users Guideparent

Comments

Here's that fun example written in Rebol parse:

    digits:     charset "1234567890"
    char:       charset [#"a" - #"z" "+/-*"]  ;; chars allowed in a symbol
    not-quot:   complement charset {"}
    esc-string: [{\"} | not-quot]

    bool:   ["#t" | "#f"]
    float:  [opt "+-" any digits "." some digits]
    number: [opt "+-" some digits]
    symbol: [some char]
    quot:   [#"'" [sexp | atom]]
    string: [{"} any esc-string {"}]

    statement: [sexp | atom | space | newline]
    sexp: ["(" any statement ")"]
    atom: [bool | float | number | quot | symbol | string]

    lisp-rule: [some statement]
NB. I wrote this "off the cuff" so I don't expect it to be flawless! Also there are no capture rules provided in my example (unlike perl6 grammars which provides this automatically).

It's worth noting that this grammar is a bit more restrictive than the Perl 6 equivalent. In particular, "\d" in Perl includes not only 0 through 9, but any Unicode character that's classified as a "digit". The Perl 6 grammar documentation features this in an example of using arbitrary methods in grammars: https://docs.perl6.org/language/grammars.html#Methods_in_Gra...

I think the Perl 6 example should really use [0-9] instead of \d because it's unlikely that unicode numerals are allowed (in the Common Lisp spec).

Anyway neither example fully covers the Common Lisp spec for a number atom (for eg. scientific notation or rationals).

I don't think covering all of the Common Lisp spec is required to be Greenspun-compliant ;)

In that case Rebol is already Greenspun-compliant ;-)

    (print (multiply (add 1 2) 3))
... will print an answer of 9
AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.