Bug in Wp-method

[Jaxan]

As far as I understand the paper of Fujiwara (and other relevant literature), the Wp-method works as follows:

1. We first test states: state prefix * middle part * suffix, where suffix are drawn from a global set.
2. Secondly we test transitions: transition prefix * middle * suffix, where the suffix is draw from a set which is relevant for a specific state, namely the state after applying transition prefix * middle. But in the implementation the state being used is the one after applying only transition prefix, as seen here: line 131

[misberner]

This is entirely correct, I will change the implementation of the Wp method accordingly.