The previous implementation did not use the first slot of the coverage
map. The first edge ID returned was 1.

and a memory-sync implementation. The implemented policy is a
refactoring of the current edge ID generation approach that should be
used when fuzzing using a single process.

This removes duplicate type definitions and makes it clear when calling
functions from the native addon from the other packages.

This way, these functions will have access to `EdgeIdStrategy` object

This strategy is only used for testing purposes

This is a first step to simplify the interface to the instrumentor
package and makes testing easier.