Endless delays and resumes.

From: Theofilos Tsoris <theofilos_at_lcn.nl>
Date: Tue 15 Sep 1998 11:57:40 AM GMT
Message-Id: <E0zJaX0-0001Mx-00@passion.icparc.ic.ac.uk>
Hi everybody,
I encountered a strange behaviour of ECLiPSe, while running the 
following procedure:

test9(X, B, Y) :-
    B::0..1,
    Y::0..1,
        #\/(X #= 1, Y #=1, B).

Now, if you run it with:
[eclipse]: test9(X,B,Y), Y=0.

it gets stuck. 
If you debug it, you will notice that the constraint is being 
delayed, since no variable is known
(eg.   (9) 1  DELAY  #>=(_227{[0, 1]} + _310{[0, 1]}, 1, B{[0, 1]})
          (8) 1  EXIT   #>=(_227{[0, 1]} + _310{[0, 1]}+-1, 0, B{[0, 
1]})  (dbg)?- creep ) 

and the procedure exits 
(eg.   (1) 0  EXIT   test9(X, B{[0, 1]},  Y{[0, 1]}) (dbg)?- creep) 
But when Y gets the value 0, the constrain is resumed, but
it can not be evaluated still
(eg.   (9) 0  RESUME #>=(_227{[0, 1]}+0, 1, B{[0, 1]}) (dbg)?- creep
          (9) 0  EXIT   #>=(_227{[0, 1]}+0, 1, B{[0, 1]}) (dbg)?- 
creep )

This resuming is continued endlessly.

I wonder if this is a normal behaviour of ECLiPSe.

Greetings,
Theofilos Tsoris



[ Moderator's note:
  A fix for this problem is now available from the ECLiPSe web site at
  http://www.icparc.ic.ac.uk/eclipse/download/release/4.0/zpatch.tgz
]
Received on Thu Sep 17 10:42:23 1998

This archive was generated by hypermail 2.1.8 : Wed 16 Nov 2005 06:07:03 PM GMT GMT