Philip McGrath writes: > Alternatively, here is a little example of how you might define lambda-box: > > #lang racket > > (struct lambda-box ([proc #:mutable]) > #:property prop:procedure > (make-keyword-procedure > (λ (kws kw-args this . by-pos-args) > (keyword-apply (lambda-box-proc this) > kws > kw-args > by-pos-args)))) > > (define f > (lambda-box (λ () 1))) > > (f) > > (set-lambda-box-proc! f (λ () 2)) > > (f)
Thanks! This is a very helpful. I guess if used in combination with namespace-variable-value one could extend or freshly define depending on whether or not things are defined in (current-namespace). > I guess my bigger question is whether making everything redefinable is > really the best mechanism to achieve what you want. I would bet that, even > in your example, you don't usually want to dynamically redefine cons or +. > I think it would probably be better to be explicit about where you want to > allow redefinition and where you want identifiers to refer to a specific > binding. > > If I were thinking about "a multiplayer game … where you could change the > world while players are in it without kicking them out," I might think > about a current-world parameter that can be mutated and having certain > functions consult the current-world. Using the class-based object system or > racket/generic for the world value might be particularly suitable. (Of > course, the details would depend on exactly what about the world you want > to be able to change.) It could be that there are per-instance ways to handle things. I'm partly coming from the perspective of other lisps where there's a very *general purpose* option with a mutable toplevel. It could be that in various applications there may be a more *specific* options that are better. But a lot of my question is whether or not the general route *can* be brought to Racket. It looks like it can be, even if Racket folks will in general not like it :) > -Philip -- 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]. For more options, visit https://groups.google.com/d/optout.
