Current coverage highlights are: "r" ran "m" missing "x" excluded (never seen yet) "p" (unknown, never seen yet)
This makes perfect sense when all tests pass.
Now consider one or a few tests fail, wouldn't it be nice to mark each "ran" line as one of these categories: ran, all relevant tests passed ran, all relevant tests failed * mixed
I'd gladly use this functionality to trace down changes in someone else's code that broke somebody else's tests. It's faster than going line by line in pdb, right?