This is not really a bug fix, but rather a suggestion for possible code improvement... Please feel free to dismiss.
Why are single precision floating point integers used only sometimes? I have found at least one case where the previous (mixed single and double precision) workflow causes errors in mesh adaptation that are not encountered after this change.
This is not really a bug fix, but rather a suggestion for possible code improvement... Please feel free to dismiss. Why are single precision floating point integers used only sometimes? I have found at least one case where the previous (mixed single and double precision) workflow causes errors in mesh adaptation that are not encountered after this change.