Possible bug within GNU Prolog ?

[Vic Bancroft]

The second clause in the predicate occured_once/2 relies on the behavior of member/2 in that if the list contains an unbound variable, it counts as an occurrance, e.g.,

  | ?- member( 1, [2,X]).
 
  X = 1 ?
 
  yes

If an unbound variable is an element of the list, then the variable can unify with operand one and occurs, e.g.,

  | ?- occured_once(1, [V]).
 
  V = 1 ?
 
  yes


If more than one unbound variable is an element of the list, then the occured_once predicate will always fail, e.g.,

  | ?- occured_once(1, [V,W]).
 
  no

The example is even more interesting when using structures,

  | ?- occured_once( v(c(1)), [ v(2), v(c(X)) ]).
 
  X = 1 ?
 
  yes


more,
l8r,
v