Comment 5 for bug 1576690

^^^ says:

16:05:00 12: [ debug ] 98 frames sampled, averaging 2.6 frames latency
16:05:00 12: [ debug ] 0: 1 2 2 2 2 2 2 2 2 2
16:05:00 12: [ debug ] 10: 2 2 2 2 2 2 2 2 2 2
16:05:00 12: [ debug ] 20: 2 2 2 2 2 2 2 2 2 3
16:05:00 12: [ debug ] 30: 3 3 2 2 2 2 2 2 3 3
16:05:00 12: [ debug ] 40: 3 3 3 3 3 3 3 3 3 3
16:05:00 12: [ debug ] 50: 3 3 3 3 3 3 3 3 3 3
16:05:00 12: [ debug ] 60: 3 3 3 3 3 3 3 3 3 3
16:05:00 12: [ debug ] 70: 3 3 3 3 3 3 3 3 3 3
16:05:00 12: [ debug ] 80: 3 3 3 3 3 3 3 3 3 3
16:05:00 12: [ debug ] 90: 3 3 3 3 3 3 3 3
16:05:00 12: /��BUILDDIR��/mir-0.23.0+vivid1121bzr3512/tests/acceptance-tests/test_latency.cpp:296: Failure
16:05:00 12: Value of: observed_latency
16:05:00 12: Expected: is < 2.4
16:05:00 12: Actual: 2.63265 (of type float)
16:05:00 12: [2016-05-19 16:05:00.057651] mirserver: Stopping
16:05:00 12: [ FAILED ] ClientLatency.triple_buffered_client_has_less_than_two_frames_latency (1821 ms)

It's almost suspicious we're missing the mark by 0.2 frames each time. Such a small amount...