Intermittent browser_bookmark_popup.js | Test timed out | Found a tab after previous test timed out: about:home

RESOLVED FIXED

Status

()

defect
RESOLVED FIXED
3 years ago
3 years ago

People

(Reporter: KWierso, Assigned: jaws)

Tracking

({intermittent-failure})

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

Comment hidden (Intermittent Failures Robot)
Blocks: 1219794
Flags: needinfo?(jaws)
I think we should disable this test for Linux debug.

There seems to be an issue with getting this mouseover event on Linux debug builds. I was seeing this while writing the test and thought I had it fixed when I landed, which is why there is a good amount of logging around these mouseover/mouseout events in the test. We would still have coverage on Linux opt builds, as well as OSX and Windows.
Flags: needinfo?(jaws)
Assignee: nobody → jaws
Status: NEW → ASSIGNED

Comment 4

3 years ago
Comment on attachment 8723400 [details]
MozReview Request: Bug 1250732 - Disable browser_bookmark_popup.js on Linux debug builds due to flakiness of moueover/mouseout events. r?gijs

https://reviewboard.mozilla.org/r/36525/#review33109
Attachment #8723400 - Flags: review?(gijskruitbosch+bugs) → review+
landed this directly on m-c in  https://hg.mozilla.org/mozilla-central/rev/4f64ec5486f4 so that this get merged to m-i/fx-team in the next merge since this is a frequent test failures
Status: ASSIGNED → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → FIXED
Not sure why I didn't think of this last night, but we could have just skipped this specific subtest `if (AppConstants.platform == "linux" && AppConstants.DEBUG)`
Comment hidden (Intermittent Failures Robot)
Gijs, do you think this could be failing if the cursor happens to already be over the future location of the bookmarksPanel when the window is created, and thus `EventUtils.synthesizeNativeMouseMove(bookmarkPanel, ...);` doesn't actually generate a new mouseover event?

We could synthesize a mousemove to the back-button at the beginning of the test as a workaround.
Flags: needinfo?(gijskruitbosch+bugs)

Comment 9

3 years ago
(In reply to Jared Wein [:jaws] (please needinfo? me) from comment #8)
> Gijs, do you think this could be failing if the cursor happens to already be
> over the future location of the bookmarksPanel when the window is created,
> and thus `EventUtils.synthesizeNativeMouseMove(bookmarkPanel, ...);` doesn't
> actually generate a new mouseover event?

I don't know. It sounds... possible but unlikely? Are there other explanations?

OTOH, see the cluster... that was bug 1205620 comment #0 / https://twitter.com/gijskruitbosch/status/644417679923761152 . So, totally possible!

FWIW, I thought native mouse events were not really native on Linux anyway, so I'm not sure how much you could depend on them in the first place, but it seems windows + osx failures also occurred...

> We could synthesize a mousemove to the back-button at the beginning of the
> test as a workaround.

We could, but we should probably make sure we leave the mouse in the same position at the end of the test as it was beforehand.
Flags: needinfo?(gijskruitbosch+bugs)
Comment hidden (Intermittent Failures Robot)
You need to log in before you can comment on or make changes to this bug.