Playing with this program. Its output is in the comments. ; /tmp/gen.rkt:9:10: assertion violation ; expected: a procedure ; given: #f ; in: method m ; (I/c (m (-> I? 1))) ; [...removed the rest of the error...] ; 1
(module gen-contracts racket/base (require racket/contract racket/exn racket/generic) (define-generics I [m I] #:fallbacks [(define (m _) 1)]) (struct noop () #:methods gen:I []) (struct impl () #:methods gen:I [(define (m _) 1)]) (define (try ctor) (displayln (with-handlers ([exn:fail:contract? exn->string]) (m (invariant-assertion (I/c [m (-> I? 1)]) (ctor)))))) (try noop) (try impl)) I expected the output to be just two "1" lines, because I assumed the contract combinator from `define-generics` would consider fallback implementations. I do think it makes sense for `I/c` to fail to apply a contract to a method in `noop` because `noop` doesn't implement `m`. But I think the combinator reuses some logic from whatever is making `#:defined-predicate` work, since it represents missing methods as `#f`, and doesn't look at the fallbacks. But... I wanted a contract that's okay with a missing method if there's a fallback. The only contracts I can write with the `I/c` combinator seem require me to either A) write repetitive method definitions that just parrot the fallback, or B) write its method contracts like `(I/c [m (or/c #f (-> I? 1))])`, which feels icky. Apologies if I'm totally off the mark or just missed something in the manual, but is there a way to define a contract for a generics implementation that won't complain if it sees a fallback for a missing method? -- ~slg -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/fed6c561-881b-08af-4318-25750f198133%40sagegerard.com.
