Graphs, graphs, graphs

So I will be honest: I didn't get much work done over Hendricks.

But I did get some interesting stuff done, so I will go over that.

I went made a racket-syntax-object to graphviz coverter:

#lang racket
#|
dotgraph.rkt

TO-USE:
    This script is currently not in an optimal state to
    be used, but if you must:
        1. Generate a datum using the BRAG library (http://docs.racket-lang.org/brag/)
        2. Change the second to last expression starting with "resolve-node")
        3. copy and paste the resulting output into a graphviz generator (IE http://cs.marlboro.edu//on-campus/graphviz/dot.cgi)
TODO:
    Create an actual interface for this.

10/18/16 | Racket 6.3 | MIT License
|#
(define dot-graph "")
(define node-number 0)

(define resolve-node
  #| Consumes a datum of a syntax object and prints
     a graphviz spec for the ast.
   |#
  (lambda (datnum)
    (define node-name (number->string node-number))
    (define grammar-rule (symbol->string (first datnum)))
    (define raw-content (rest datnum))
    (define named-content '())
    (set! node-number (+ node-number 1))
    (for ([item raw-content])
      (cond  [(list? item) (and (set! named-content (append named-content (list (list (number->string node-number) item))))
                                (resolve-node item))]
             [(string? item) (set! named-content (append named-content (list (list (number->string node-number) item))))]
             [else (set! named-content (append named-content (list (list (number->string node-number) (symbol->string item)))))])
      (set! node-number (+ node-number 1)))
    (for ([item named-content])
      (set! dot-graph (string-append dot-graph node-name " -> " (first item) ";\n"
                                     (first item)"[label=\"" (~a (second item)) "\"]" ";\n")))))

(resolve-node '(program (lang-name "#lang" "racket") (s-exp "(" "+" "6" (s-exp "(" "-" "7" "8" ")") ")"))) ;; edit this thing
(printf "~a" dot-graph)

The resulting spec was fed into this page: http://cs.marlboro.edu//on-campus/graphviz/dot.cgi

I used the same simple test program with two different grammars to see the difference in parsing. Here is the sample program:

#lang racket
(+ 6 (- 7 8))

This is the simpler of the two grammars I used:


#lang brag

program: lang-name s-exp*
s-exp: LPAREN (KEYWORD | ID | s-exp| QUOTE s-exp | LIT | SYM | COMMENT)* RPAREN
lang-name: META-WORD ID

Which produces this graph:

The other grammar was this:


#lang brag

program: lang-name (s-exp| comment)*
s-exp: (list | quoted-list)+
quoted-list: QUOTE list
list: LPAREN (list | quoted-list | single-value)* RPAREN
single-value: (data | symbol)
data: (LIT | unresolved-name)
unresolved-name: (KEYWORD | ID)
symbol: SYM
comment: COMMENT
lang-name: META-WORD ID

Which made this graph: