module Partitioning_index:sig
..end
A partitioning index is a collection of states optimized to determine if a new state is included in one of the states it contains — in a more efficient way than to test the inclusion with all stored states. Such an index is used to keep track of all the states already propagated through a control point, and to rule out new incoming states included in previous ones.
Partitioning index relies on an heuristics on the cvalue domain, and is very inefficient without it.
module type Domain =sig
..end
module Make:
Partition of the abstract states, computed for each node by the dataflow analysis.