[MLton-user] opaque types and ffi

[Matthew Fluet]

> Hello, I'm creating a structure for ffi pointers. The idea is to use
> the word type and remove almost all operations using an opaque signature
> match. However, the opaque match seems to destroy the ability to pass
> the ffi. Is it possible to write such a structure?
>
> This is more like a theoretical question since it's not really necessary
> to use :> instead of :. Vesa

The short answer is no.  Once you use an opaque signature match, the type
becomes abstract and the type checker will refuse to pass an abstract type
to an ffi function.