Spurious answers sets if there are theory atoms with trivially false bodies

[rkaminsk]

If a program contains a theory atom with a trivially false body, then there are too many answer sets. I suspect that clasp treats it like a theory atom in a rule body in this case:

#theory bug {
  t {}; 
  &x/0: t,head       
}.                            
                                     
{c}.
:- c.    
&x { c } :- c.

Hiding the false body a bit produces the right results:

```prolog
#theory bug {
  t {}; 
  &x/0: t,head       
}.                            
                                     
{c,d}.
:- c, d.    
:- c, not d.    
:- not c, d.    
&x { c } :- c.

The aspif of the problematic version is:

asp 1 0 0
1 1 1 1 0 0
1 0 1 2 0 1 1
1 0 0 0 1 1
9 1 0 1 x
9 1 1 1 c
9 4 0 1 1 0
9 5 2 0 1 0
4 1 c 1 1
0

And the output of clasp is:

clasp version 3.3.6
Reading from stdin
Solving...
Answer: 1

Answer: 2

SATISFIABLE

Models       : 2
Calls        : 1
Time         : 0.000s (Solving: 0.00s 1st Model: 0.00s Unsat: 0.00s)
CPU Time     : 0.000s

[BenKaufmann]

@rkaminsk
Pushed a fix to dev.