Handle cases where test executor does not return "label"

[sven-oly]

This test driver and the verifier should give information on executor failures in which individual tests are not returned.

Current behavior on a serious executor problem results in many missing tests.

[sffc]

This is important for overall integrity of the dashboard because missing tests may or may not be noticed. They should be regarded as errors or something. Would be OK marking them with a new color. Maybe purple or brown.

[sffc]

I think missing tests should show up in the results; should they be classified as failures? Or maybe a new category like missing (black)?