Asumu Takikawa writes:
> This is what the `define/generic` form is for. You can then re-write
> your struct definition as
Great, thanks! I had seen define/generic but its documentation made me
believe that its use is more restricted. But it works fine in my
demonstration code.
Unfortunately, I did not manage to fix the real code that motivated my
question using the same trick. The difference is that in the real code,
the method name shadows not a local function, but a field accessor in
the very structure that I am defining the methods for.
Here's an updated version of my demo code:
-- foo.rkt ------------------------------------------------
#lang racket
(provide (all-defined-out))
(require racket/generic)
(define-generics bar
[foo-bar-baz bar])
-----------------------------------------------------------
-- foo-bar.rkt --------------------------------------------
#lang racket
(provide (all-defined-out))
(require (prefix-in foo: "foo.rkt")
(only-in racket/generic define/generic))
(struct foo-bar (baz)
#:methods foo:gen:bar
[(define/generic gen-foo-bar-baz foo-bar-baz)
(define (foo-bar-baz x)
(gen-foo-bar-baz x))])
-----------------------------------------------------------
Then, in a REPL:
(require (prefix-in foo: "foo.rkt")
"foo-bar.rkt")
(foo:foo-bar-baz (foo-bar 42))
This hangs in the method foo-bar-baz, so it looks like my
freshly defined gen-foo-bar-baz still refers to the method
being defined.
Is there a workaround for that case as well?
Thanks,
Konrad.
____________________
Racket Users list:
http://lists.racket-lang.org/users
