More precise float domain for `float` and `double`

[michael-schwarz]

In these cases we can have a more precise interval domain (float option * float option) option * bool where the inner values are for the interval (None for the infinite cases), and the outer bool represents whether NaN is a possible value or not.

Currently we cannot benefit from guards such if(5.0f < x && x < 10.0f): For the first part, we cannot express that x needs to be betweeen 5 and +Infinity, as our domain expresses this as T.

Such a treatment is only possible when we know that the type we are using to represent a value matches the concrete one. If we will participate in SV-COMP 24, that might yield some improvement in the floats category.