If I have a NaBL rule like this:

TypeDef(name, cnt):
	defines Type name
	of param-count cnt

Then two TypeDefs that only differ on param-count, e.g. TypeDef("Func", 1) and TypeDef("Func", 2), result in a Duplicate definition error. I would expect NaBL to distinguish the two definitions by their properties and not generate an error. This is neccesary to implement generics (distinguishing Func<TResult> from Func<T1, TResult>) and overloading (distinguishing f(x : int) from f(x : string)).

I’ve added a project that demonstrates this issue.

I can’t use the non-unique keyword here, as that would join all definitions into one, whereas I need separate definitions.

Submitted by D. Pelsmaeker on 27 June 2014 at 12:26

You should turn off the default duplicate check and implement your own check instead. See the assignments of IN4303 on how to do this.

This currently is intended behaviour. You can disable the duplicate definition error by adding nabl-check-disable-duplicate(|uri, ns) = ?NablNsType(), and creating your own constraint rule that checks for this.