It turns out that the continuation mark keys and procedures for parameters aren’t really hidden away, so it ended up being pretty straight forward. Below is a minimal implementation of parameterization->. I’m happy to help integrate this into the racket/contract family once we figure out what’s the nicest interface.
#lang racket (require '#%paramz) (define (parameterization-> param-name param val/c) (make-chaperone-contract #:name `(parameterization-> ,param-name ,(contract-name val/c)) #:first-order procedure? #:late-neg-projection (λ (blame) (λ (value party) (let* ([positive (blame-replace-negative blame party)] [negative (blame-swap positive)]) (unless (procedure? value) (raise-blame-error positive value '(expected: "procedure?" given: "~e") value)) (chaperone-procedure value (λ args (let ([new-parameterization (extend-parameterization (continuation-mark-set-first #f parameterization-key) param (((contract-projection val/c) negative) (param)))]) (apply values 'mark parameterization-key new-parameterization args))))))))) (require rackunit) (define p (make-parameter #f)) (define/contract (foo) (parameterization-> 'p p integer?) (p)) (define/contract (bar) (parameterization-> 'p p (-> integer? integer?)) ((p) 4)) (define/contract (baz) (parameterization-> 'p p (-> symbol? symbol?)) ((p) 4)) (check-exn exn:fail:contract? (thunk (foo))) (check-not-exn (thunk (parameterize ([p 2]) (foo)))) (check-not-exn (thunk (parameterize ([p (λ (x) x)]) (bar)))) (check-exn exn:fail:contract? (thunk (parameterize ([p (λ (x) x)]) (baz)))) On November 28, 2016 at 11:24:29 AM, Alexis King (lexi.lam...@gmail.com) wrote: That sounds promising, yes. Not being familiar with the guts of parameters, is there any way to implement this as a derived concept using the existing support in chaperone-procedure? As far as I can tell, parameters do not expose the continuation marks they use, and they also create thread cells, which I’m not sure that chaperone-procedure’s existing API would support. Would this require modification of procedure chaperones to support parameters directly, or is there some way to implement it separately? > On Nov 23, 2016, at 10:51 AM, Scott Moore <sdmo...@fas.harvard.edu> wrote: > > Yes, we worked with Matthew to implement the necessary hooks in procedure > chaperones (see the 'mark options that were added to the return value of > wrapper-proc). For the contracts we were writing, we ended up using these > continuation marks directly. > > To implement what you're looking for, a little extra work is required to link > up the implementation of parameters with this mechanism to get at the > appropriate continuation marks. One question there will be whether to > integrate it with either the existing > arrow contracts (carefully protecting access to the internals of the > parameter implementation), or to just provide a standalone combinator. I > would need to refresh my memory to see what exactly would need to be done for > either. > > Cheers, > Scott -- 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. For more options, visit https://groups.google.com/d/optout.
