[Quant] Substitute by prefering existing terms to new equal one

See merge request !25

    since 0 added to sign

      It is complicated to know who is responsible for this choice

       same for ceiling, floor, floor_to_int

        Help to keep synchronized the information about types and the events

        Avoid useless reasonning:
          - congruence closure
          - simplex
          - instantiations

 Reove the hack about coercions that was just removing problematic axioms

Release helpers

See merge request !24