fd library

Dear Sir, Madam

We have noticed an abnormal behavior of the logical constraint
predicate   #/\   of fd library.
____________________________________________________________________________

ECLiPSe Constraint Logic Programming System [sepia parallel mps]
Version 4.0.2, Copyright ECRC GmbH and ICL/IC-Parc, Fri Jul 28 15:30
1998

[eclipse 1]: lib(fd).
structures.pl compiled traceable 3984 bytes in 0.00 seconds
lists.pl   compiled traceable 6436 bytes in 0.01 seconds
sorts.pl   compiled traceable 1292 bytes in 0.00 seconds
fd_domain.sd loaded traceable 27392 bytes in 0.05 seconds
fd_arith.sd loaded traceable 72912 bytes in 0.13 seconds
fd_util.pl compiled traceable 2116 bytes in 0.01 seconds
fd_chip.pl compiled traceable 4744 bytes in 0.02 seconds
fd_elipsys.pl compiled traceable 10836 bytes in 0.02 seconds
fd.sd      loaded traceable 19300 bytes in 0.20 seconds

yes.
[eclipse 2]: [X,Z] :: -100..100, Z#= X, X  #> 0.

Z = X{[1..100]}
X = X{[1..100]}
yes.
[eclipse 3]: [X,Z] :: -100..100, (Z#= X #/\  X  #> 0).

no (more) solution.
[eclipse 4]: [X,Z] :: -100..100,(Z#= -X #/\  X  #> 0).

Z = Z{[-100..-1]}
X = X{[1..100]}

Delayed goals:
        0 + X{[1..100]} + Z{[-100..-1]}#=0
yes.
[eclipse 5]: [X,Z] :: -100..100,(Z#= X + 0 #/\  X  #> 0).

Z = X{[1..100]}
X = X{[1..100]}
yes.

________________________________________________________________________________

Is this behaviour justified? If yes, how?
If not, is there a more recent version of the fd library?
Sincerely,

Christiane Hespel
IRISA-INSA
20 avenue des Buttes de Coesmes
35043 Rennes cedex
France