[mercury-users] Indexing & operational semantics

[Ralph Becket]

I'm pretty sure the answer to this question is `yes', but
I just want to check.

I'm writing a term -> my_thing translator and I want to
raise an error if it sees a syntax error or attempt to
overload a primitive functor name.  Some of the clauses
in the translator therefore look like this:

translate(functor(atom("/\"), Args, Ctxt), Out) :-
	require(list__length(Args) = 2, "/\ must have two args"),
	Out = f(Args).

Now it occurred to me that this could equally be written
as

translate(functor(Name, Args, Ctxt), Out) :-
	Name = atom("/\"),
	require(list__length(Args) = 2, "/\ must have two args"),
	Out = f(Args).

which is something the compiler might very well do.  In this
case, there's no dependency between the first and second lines
of the clause body, so there's a chance that they will be
reordered.  This would be a disaster, since I only want require/2
to be called provided Name = atom("/\").  Since this is the
behaviour that I assume everybody expects, I infer that head
unifications work like this:

translate(functor(Name, Args, Ctxt), Out) :-
	( if Name = atom("/\") then
		require(list__length(Args) = 2, "/\ must have two args"),
		Out = f(Args)
	  else
		fail
	).

with appropriate munging of output variables from the head 
unification.  I just couldn't find anything about this in the
reference manual.

Ralph
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------