Intermittent browser_copypanel.js | Test timed out -

RESOLVED INCOMPLETE

Status

defect
P1
normal
Rank:
22
RESOLVED INCOMPLETE
3 years ago
2 years ago

People

(Reporter: cbook, Assigned: Mardak)

Tracking

({intermittent-failure})

unspecified
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(firefox47 unaffected, firefox48 affected, firefox49 disabled)

Details

(Whiteboard: [btpp-fix-now][test disabled on linux], )

Attachments

(2 attachments)

https://treeherder.mozilla.org/logviewer.html#?job_id=8826234&repo=fx-team

 07:42:10 INFO - 320 INFO TEST-UNEXPECTED-FAIL | browser/extensions/loop/chrome/test/mochitest/browser_copypanel.js | Test timed out -
Ed: can you look at this one?
Rank: 19
Flags: needinfo?(edilee)
Priority: -- → P1
Whiteboard: [btpp-fix-now]
(In reply to OrangeFactor Robot from comment #5)
> 29 automation job failures were associated with this bug yesterday.
Hrm.. this seems like quite the spike compared to less than 2 a day from the weeks before.

There is bug 1273671 that landed 1.4.0 in m-c, but nothing obvious about what might have caused the spike. I'll try to poke around with my linux vm.

Standard8, any initial thoughts on what might have caused the sudden increase?
Assignee: nobody → edilee
Flags: needinfo?(edilee) → needinfo?(standard8)
I've had a look through some of the diffs, and I can't see anything obvious, unless some of the babel or eslint changes messed us up.
Flags: needinfo?(standard8)
#1 on OrangeFactor by a mile now. Anybody up for pointing rr-chaos at it? :)
Flags: needinfo?(edilee)
I've been doing back-to-back runs on my Linux VM with low memory ~336MB, but it hasn't timed out. If I lower the memory too much, it starts complaining that the test should be split up.

Any specific instructions for using rr debugging add-on js code? I'll start digging through http://rr-project.org/
I took another brief look at the code having seen the comments here.

There's one piece in testClick where the code is invoking cmd_copy before adding a popupshown listener. My guess is it could be that for some reason on Linux the listener doesn't always get added before the cmd_copy has completed and has shown the panel.

I could be wrong, but I'm semi-basing this on the fact that it seems to be the Linux opt/pgo builds that are the ones failing, not the debug ones (hence faster test builders potentially the ones at issue).

In any case, I've pushed adjusting the listener flow to try, to see if it helps or not:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=afa9dbfc844c

I've also just noticed there's waitForLoopPanelShowHide which is adding listeners and looks like it could be a bit suspect, but lets see what the builders say (if they fail, we need to check exactly which bit of test is failing).
That didn't seem to work :-(

Ed, have you tried adding various debug prints, to see if we can work out if the click is happening/where exactly we're hanging?
Haven't been able to figure out what caused the spike or had much time to debug this leading up to london. We'll just disable this for now.
Flags: needinfo?(edilee)
Attachment #8759264 - Flags: review?(dmose)
Comment on attachment 8759260 [details] [review]
[loop] Mardak:bug-1266041-disable-test > mozilla:master

Same patch but for loop repository.
Attachment #8759260 - Flags: review?(dmose)
Comment on attachment 8759264 [details] [diff] [review]
v1 disable test on linux m-c patch

Review of attachment 8759264 [details] [diff] [review]:
-----------------------------------------------------------------

rs=dmose
Attachment #8759264 - Flags: review?(dmose) → review+
Comment on attachment 8759260 [details] [review]
[loop] Mardak:bug-1266041-disable-test > mozilla:master

rs=dmose
Attachment #8759260 - Flags: review?(dmose) → review+
Whiteboard: [btpp-fix-now] → [btpp-fix-now][test disabled on linux][checkin-needed-48]
https://hg.mozilla.org/releases/mozilla-aurora/rev/072ce9419eaa
Whiteboard: [btpp-fix-now][test disabled on linux][checkin-needed-48] → [btpp-fix-now][test disabled on linux]
Rank: 19 → 22
Support for Hello/Loop has been discontinued.

https://support.mozilla.org/kb/hello-status

Hence closing the old bugs. Thank you for your support.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INCOMPLETE
Removing leave-open keyword from resolved bugs, per :sylvestre.
Keywords: leave-open
You need to log in before you can comment on or make changes to this bug.