The implementation achieves domain consistency iff C is instantiated at call time, otherwise only bounds consistency.
?- [X,Y] &:: weekday, shift(X, 1, Y).
X = X{[mo, tu, we, th, fr, sa]}
Y = Y{[tu, we, th, fr, sa, su]}
There is 1 delayed goal.
Yes (0.00s cpu)
[eclipse 4]: [X,Y]&::weekday, shift(X,C,Y).
X = X{[mo, tu, we, th, fr, sa, su]}
C = C{-6 .. 6}
Y = Y{[mo, tu, we, th, fr, sa, su]}
There are 3 delayed goals.
Yes (0.00s cpu)
?- shift(we, 1, th).
Yes (0.00s cpu)
?- shift(we, 2, fr).
Yes (0.00s cpu)
?- shift(X, -1, th).
X = fr
Yes (0.00s cpu)
?- shift(tu, X, fr).
X = 3
Yes (0.00s cpu)
?- shift(tu,X,Y).
X = X{-1 .. 5}
Y = Y{[mo, tu, we, th, fr, sa, su]}
Delayed goals: ...
?- shift(tu, 1, th).
No (0.00s cpu)
?- shift(X, 1, Y).
Arguments have no domains in shift(X, 1, Y) in module eclipse
Abort
?- X &:: weekday, shift(X, 1, red).
Arguments have different domains (weekday,colour) in shift(X, 1, red) ...
Abort