I'm not clear what you mean by this:
"Assigning names is an issue and I do not believe that they ignore the ethnicity in what is determined to be your direct line (as opposed to who you match)."
I do not think that is the case. The computer first identifies your matches based on segment analysis. Let's assume it's you and me. So jbarry6899 would appear on your list, and swcbcm would appear on mine. That's the first step. Then the computer pulls up both of our trees. For me, it lists on the left hand side all of the surnames in your direct lines out to 10 generations. For you, it provides a similar list of surnames from my tree. That's the second step. Let's assume that we both have the surname Costello in our trees. That would be listed at the top for both of us as a common surname. Then let's assume that we both have a great grandfather named Lou Costello in our trees. That would be flagged with a hint leaf and we could then communicate to see if it is in fact the same person. The third step. If so, we would have a confirmed match. The final step.
There is nothing in this process about ethnicity. Once again, the determination of our ethnic breakdown is an entirely separate analysis based on the relationship between your or my DNA and the reference populations, not on segment matching. So we could (and probably do) have very different ethnicity percentages, even if we both are descended from Lou Costello. Conversely, we could have identical ethnicity percentages and no common ancestors. Completely different analyses with different objectives and outputs.