Explaining causality errors by incremental solving about bonsai HOT 2 OPEN

There is another problem with the current technique: models generated share a part of their constraints.
This is for example the case for P ; when C then Q else R, a model is generated for P,C,Q and for P,!C,R and thus have P in common.
Hence, if a constraint in P is invalid, it generates two errors.
For now, to avoid this problem, we stop the causality analysis after the first error.

from bonsai.

We already perform a postfix addition of the constraints into the CSP model. Therefore, if we solve the sub-problems, we will first test the smallest block of code first, and if they are satisfiable, we test their combinations.
In addition, if we attach the partially solved CSP to each process, we also solve #5.

from bonsai.