Improper detection and marking of "new test failure"s

Take the scenario of three consecutive builds:

1. test A fails
2. compilation failure
3. test A fails

In build #3 test A is marked as a "new failure" because it didn't fail in build #2. However, there weren't any successes between failures.
If you go to the test case history it correctly shows that it has been failing since #1 (or earlier) but it would be nice if the build #3 page also said it was an existing failure. Would make it a bit easier in diagnosing/fixing broken builds.