disjoint2/1 succeeds with negative widths and heights

[SaGagnon]

Hi,

First of all, thanks a lot for this amazing library.

While using and testing the predicate disjoint/2, I noticed that the following query succeeds:

?- disjoint2([r(1,1,1,1), r(2,-2,2,-2)]).
true.

I see two problems:

1. It could be argued that negative widths and heights are possible, but I think it is confusing. Also, the Global Constraint Catalog constrains sizes to be >=0 for diffn.
2. Otherwise, if we allow negative widths and heights, I think it would be more intuitive if this query fails.

If you agree with one of the two points, I would like to try to make a contribution.

[triska]

Thank you a lot for your kind words, and for your interest in this library!

Could you please test the query with a version of SICStus Prolog (an evaluation version is available at no cost), and see what it says? I think it would be good to ensure compatibility of this constraint with SICStus Prolog. Thank you a lot!

[UWN]

SICStus succeeds. Obsolescent.

[SaGagnon]

Sadly, I already used my free trial for SICStus Prolog. In light of what UWN said, do you still want a correction or do you want to preserve the same behavior as SICStus?