Previously in some edge cases we could end up with a domain that is not of the correct mesh size. In this PR we first round so we compensate for floating point inaccuracies and we also add an assertion to make sure the domain has the correct mesh size.
Previously in some edge cases we could end up with a domain that is not of the correct mesh size. In this PR we first round so we compensate for floating point inaccuracies and we also add an assertion to make sure the domain has the correct mesh size.