Faster Pull Request analyses

A post was split to a new topic: C++ PR analysis is not faster with server side cache