[m-dev.] clauses --> switches

[Fergus Henderson]

On 29-May-1998, Erwan Jahier <Erwan.Jahier at irisa.fr> wrote:
> Let's consider a Mercury predicate defining by several clauses. 
> Is it true to say that those clauses will be converted into exactly one 
> switch ?

No.

> Are there some cases where it can be splited into several
> switches (on different arguments of the predicate) ?

Yes, either on different arguments of the predicate, or
on different parts of the same argument. E.g.

	:- type foo ---> f(char) ; g(char, char).

	:- pred p(foo, int).
	:- mode p(in, out).
	p(f(a), 1).
	p(f(b), 2).
	p(g(c,d), 3).
	p(g(c,e), 4).

This will be split into three switches.  The top-level switch will switch
between the first argument being f/1 and g/2.  The switch in the one
branch of the main switch will switch between the first argument of
the functor f/1 being a/0 and b/0.  The switch in the other branch of
the main switch will switch between the second argument of g/2
being d/0 or e/0.

> In other words, is it true to say that a goalpath of the form "s1;..."
> corresponds to an event occuring in the first clause, "s2;..." in the
> second clause, ... ?

No.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3        |     -- the last words of T. S. Garp.