Make cutsets smaller in DataLang

[myreen]

@tanyongkiam noticed that cutsets are sometimes too large.

For example, the explorer output reveals:

         (seq
            ...
            (6 := Mult (5 4) (some {4,5}))
            (return 6))

Here 4 and 5 are in the cutset ({4,5}) only because they are used as arguments to Mult. Ideally, the cutset should be empty because 4 and 5 are not used afterMult.

The reason they are in the cutset can be seen in the semantics of DataLang, here:

cakeml/compiler/backend/semantics/dataSemScript.sml

Lines 1146 to 1153 in 48a421d

	(evaluate (Assign dest op args names_opt,s) =
	if op_requires_names op /\ IS_NONE names_opt then (SOME (Rerr(Rabort Rtype_error)),s) else
	case cut_state_opt names_opt s of
	| NONE => (SOME (Rerr(Rabort Rtype_error)),s)
	| SOME s =>
	(case get_vars args s.locals of
	| NONE => (SOME (Rerr(Rabort Rtype_error)),s)
	| SOME xs => (case do_app op xs s of

This needs to be corrected so that the arguments are read (get_vars) before the cutset is applied (cut_state_opt). With this change data_live can make the cutsets smaller. The data_to_word proofs need updating.