Constraint programs such as those written in modern Con- straint Programming languages and platforms aim at solving problems coming from optimization, scheduling, planning, etc. Recently CP pro- grams have been used in business-critical or safety-critical areas as well, e.g., e-Commerce, air-traffic control applications, or software verification. This implies a more skeptical regard on the implementation of constraint solvers, especially when the result is that a constraint problem has no solution, i.e., unsatisfiability. For example, in software model checking, using an unsafe constraint solver may result in a dramatic wrong an- swer saying that a safety property is satisfied while there exist counter- examples. In this paper, we present a...