Fix running GkConj twice

[rTreutlein]

@patham9 This breaks a test but i am not sure that the way the test (spaces_find) works now is correct could you please check that.

[patham9]

Does this not change semantics? The empty case should only run if there was no item at all generated.

[patham9]

Changes semantics.

(friend a b)
!(case (match &self (friend a b) True)
       ((True True)
        (Empty False)))

WIth your version it reaches true and false both, even though the pattern is in the space.

[patham9]

Closed as this is not a solution.