Does Racket have any grammar-based fuzzing utilities? I'm probably going to
play around with this either way, but if there is an existing solution I'll
quit after I've had my fun. If, however, people think Racket could use
something like this I may attempt to make it into a usable package.
I'm thinking specifically about generating inputs for Quickcheck, but such
a fuzzer might have other uses.
For example, given this grammar:
(struct tree (left right) #:transparent)
(define-grammar tree-grammar
[Leaf (:choose-integer -9999 9999)]
[Subtree #'(tree Node Node)]
[Node #:max-nesting 10
(:weighted-choice
[1 Leaf]
[5 Subtree])])
Then (tree-grammar 'Subtree) might generate #'(tree 1 (tree 2 3)) as one of
many possibilities.
An important feature that I haven't seen yet is tracking/scoping
identifiers. There are times when we want to generate any arbitrary
identifier, and other times when we want to use an existing identifier that
we know is in scope. For example, if we were fuzzing `let` we might have
(define-grammar let-grammar
[Root #'(let ([LetId LetVal]
...)
LetBody)]
[LetId (:choose-identifier)]
; TODO ...
)
And somehow we want to say that LetBody can and should use the enclosing
LetIds, but I don't immediately see a great way to communicate that. It
might be possible to have something like (:choose-bound LetId) which would
generate a LetId that is known to be in scope.
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/racket-users/832ed76d-33f6-4f3f-b80f-da891613b6a3%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
