[MLton] sequenceNonUnit

[Matthew Fluet]

> > it does print an error/warning for sequence expressions with 
> > non-unit types that are later constrained by a signature to be unit.
> ..
> > To fix the warning, you need to put a type annotation on cleanup.  I
> > think this behavior is fine, because it would be fairly difficult (I
> > believe) to distinguish the above case from one where wind was also
> > used within the structure at non-unit type.
> 
> Could this be accomplished by delaying the warnings a bit?  Maybe not
> a good idea because then they would be out of place with other
> messages, but, to pick an extreme, it seems like delaying the warnings
> until after elaboration is done would prevent this issue.

I don't believe that fixes the issue.  For example, what warnings would 
you expect from the following:

structure S1 : 
   sig
      val f : (unit -> unit) -> unit
   end =
struct 
   fun f thunk = (thunk (); ())
end

structure S2 : 
   sig
      val f : (unit -> unit) -> unit
      val g : (unit -> int) -> unit
   end =
struct 
   fun f thunk = (thunk (); ())
   val g = f
end

structure S3 : 
   sig
      val f : (unit -> unit) -> unit
   end =
struct 
   fun f thunk = (thunk (); ())
   val () = f (fn () => 1)
end


Constraining a structure by a signature does not alter the principal type 
of the structure, so the inferred type of the expression in a sequence 
position is not altered by the signature match.