Constraint solving techniques are frequently based on constraint propagation, a technique that can be seen as a specific form of deduction. Using constraint programming languages enhanced with constraint handling rules facilities, constraint propagation can be achieved just by applying deduction rules to constraints. The automatic generation of propagation rules has been recently investigated in the particular case of finite domains, when constraint satisfaction problems are based on predefined, explicitly given constraints. Due to its interest for practical applications, several solvers have been developed during the last decade for integrating finite domains into (constraint) logic programming. A possible way of integration is implemented using a unification algorithm to compute most general solutions of constraints. In this paper, we propose a mixed approach for designing finite domain constraints solvers: it consists in using a solver based on unification to improve the generation of propagation rules.