Closed Bug 1055087 Opened 11 years ago Closed 6 years ago

Mozlog should log a smaller set of statuses for compatibility with tbpl

Categories

(Testing :: Mozbase, defect)

x86_64
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INCOMPLETE

People

(Reporter: chmanchester, Assigned: chmanchester)

References

Details

Attachments

(1 file)

See https://bugzilla.mozilla.org/show_bug.cgi?id=1050177#c7 I have something that implements this in a patch for bug 1055070. The requirement as I understand it is to log a smaller set of statuses (tbpl knows about "PASS" and "FAIL", pretty much), and to propagate a failing status from subtests to the test_end message.
Isn't the part about propagating a failing status from subtests to test_end already done by the test harnesses currently?
Assignee: nobody → cmanchester
Status: NEW → ASSIGNED
(In reply to Ahmed Kachkach [:akachkach] from comment #1) > Isn't the part about propagating a failing status from subtests to test_end > already done by the test harnesses currently? I mean only logging "TEST-PASS" and "TEST-UNEXPECTED-FAIL" in test_end.
Hmm, why do we want to propagate the status from subtests to the test_end? I'm not thrilled about that.
(but I can probably be convinced that it's OK to do this for mochitest for legacy reasons, if it's something we can remove once mozharness uses the raw logs to set the job status, which should be a high priority).
I guess this isn't about mozharness. It's a requirement of the logs as they are today that they preserve the meaning of mozharness' parsers for all the relevant cases. Anywhere they aren't should be fixed in an urgent manner. This is more about keeping the logs looking similar to the way they did before, and possibly a precautionary measure with respect to the log parsers we haven't yet discovered. Making things look like they did before in a way that reduces the amount of information they contain is pretty lame, but it's just a formatter.
We didn't have any TEST-UNEXPECTED-FAIL on the subtest level before? If it's the case then this change makes sense since it would give the "right" number of errors in mozharness (before we finally get to refactor it)

I'm guessing this bug isn't relevant any more.

Status: ASSIGNED → RESOLVED
Closed: 6 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: