This bug is to investigate if any differences in prefix paths between the build and test machines are observed. This is needed to determine if we will ever need a more involved method of dealing with prefixes. On linux, we think that the prefix path is consistent between the test and build machines but this needs to be looked at some more. And if we start gathering coverage for windows, could we expect some difficulties there also?

I think we have fixed the prefixes problem with grcov's "-p" option.