Make visual diff threshold configurable
This provides an easy way to work around the fact that visual diff results can vary from host to host. It's based on this earlier MR.
There was discussion in that previous MR about logging the different threshold used, which this approach doesn't need I think, since the user will have explicitly set the threshold themselves when executing the tests.
It also adds a section in the README about this and how to run the tests in general, which is a must!
Closes #173 (closed)
Edited by Mark Florian