It makes no sense to ask the type system to distinguish a Tree of Integers from a Tree of Symbols? Basically, I have an informal requirement (trees are homogenous) from the untyped world that I want to formalize through require/typed and check statically.
On Sat, Dec 6, 2014 at 5:40 PM, Matthias Felleisen <[email protected]> wrote: > > This makes no sense. Parametric polymorphism is a syntactic concept (think > for-all quantifier). It's sad enough that we export such things from typed > to untyped modules (for pragmatic reasons, in violation of a fundamental GT > assumption). -- Matthias > > > > > > On Dec 6, 2014, at 5:00 PM, Benjamin Greenman wrote: > > The typed racket reference shows how to require/typed a struct from an > untyped program [1]. I'd like to generalize this example to a struct with a > type parameter. > > Here is the code I would like to write. How can I get this to type check? > > #lang racket/base > > (module UNTYPED racket/base > (struct Tree (elem left right)) > (provide (struct-out Tree))) > > (module TYPED typed/racket/base > (require/typed 'UNTYPED > [#:struct (A) Tree > ([elem : A] > [left : Tree A] > [right : Tree A])])) > > [1] > http://docs.racket-lang.org/ts-reference/special-forms.html#%28form._%28%28lib._typed-racket%2Fbase-env%2Fprims..rkt%29._require%2Ftyped%29%29 > ____________________ > Racket Users list: > http://lists.racket-lang.org/users > > >
____________________ Racket Users list: http://lists.racket-lang.org/users
