Example: bug 610338
A return_code bisect was kicked off for the failing test, and it turns out that the culprit CL also caused performance regressions. The bisect duped the test failure bug against the performance regression bug.
In my opinion, return_code bisects should not dupe bugs at all:
* Perf regression bugs shouldn't be duped into return_code bisects
* return_code bisects shouldn't dupe the bugs even if there are two return_code bugs with the same culprits, because the test re-enables are tracked in two separate bugs and one re-enable may be overlooked if they are duped.
I think the relevant code is here: https://github.com/catapult-project/catapult/blob/master/dashboard/dashboard/update_bug_with_results.py#L249
Anyone want to take a look at this? I'd be happy to provide guidance/review.
Comment 1 by pras...@chromium.org
, May 23 2016