Bugzilla

Some notes/observations on the cited pretty-print tests:
1) The graphs indicate that they are tri-model.
2) Prior to the time around when my commit landed, the fastest of the three "modes" was the most common and dominated.
3) Sometime around when my commit landed (but _possibly_ somewhat before it) the slowest of the three modes started becoming noticeably more common than before.
4) Sometime very shortly after my change, the middle mode started becoming more common.
5) Sometime after my change, maybe about a day later, the faster of the modes became very uncommon and now the test is dominated by the slowest mode.

It does seem like something real has changed wrt the pretty-print test (and it looks to me like there were two distinct changes). My questions:
1) To what extent are we certain that _any_ pretty-print test changes were caused by [the set of commits](https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=3d4f500d317cd4c54e93299c01efed1c4354deb7&tochange=7c0de707e2a448dd3846c81f12ee9bae3ebca9d3?) associated with this alert?
2) To what extent are we certain that [my specific change](https://hg.mozilla.org/integration/autoland/rev/7c0de707e2a448dd3846c81f12ee9bae3ebca9d3) is responsible for any of the pretty-print test changes?

Some notes/observations on the cited pretty-print tests:
1) The graphs indicate that they are tri-model.
2) Prior to the time around when my commit landed, the fastest of the three "modes" was the most common and dominated.
3) Sometime around when my commit landed (but _possibly_ somewhat before it) the slowest of the three modes started becoming noticeably more common than before.
4) Sometime very shortly after my change, the middle mode started becoming more common.
5) Sometime after my change, maybe about a day later, the faster of the modes became very uncommon and now the test is dominated by the slowest mode.

It does seem like something real has changed wrt the pretty-print test (and it looks to me like there were two distinct changes). My questions:
1) To what extent are we certain that _any_ pretty-print test changes were caused by [the set of commits](https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=3d4f500d317cd4c54e93299c01efed1c4354deb7&tochange=7c0de707e2a448dd3846c81f12ee9bae3ebca9d3) associated with this alert?
2) To what extent are we certain that [my specific change](https://hg.mozilla.org/integration/autoland/rev/7c0de707e2a448dd3846c81f12ee9bae3ebca9d3) is responsible for any of the pretty-print test changes?