Closed Bug 1738293 Opened 3 years ago Closed 2 years ago

Multiple intermittents from AutocompleteTest

Categories

(GeckoView :: General, defect, P1)

Unspecified
All

Tracking

(firefox98 wontfix, firefox99 wontfix, firefox100 wontfix, firefox101 wontfix, firefox102 wontfix, firefox103 wontfix, firefox104 fixed)

RESOLVED FIXED
104 Branch
Tracking Status
firefox98 --- wontfix
firefox99 --- wontfix
firefox100 --- wontfix
firefox101 --- wontfix
firefox102 --- wontfix
firefox103 --- wontfix
firefox104 --- fixed

People

(Reporter: fluffyemily, Assigned: owlish)

References

(Blocks 1 open bug)

Details

(Whiteboard: [geckoview:m104])

Attachments

(1 file)

We have lots of intermittents happening in the autocomplete tests

Priority: -- → P2

I'll take a look at this since I ran into it when working on another bug and I can reproduce it locally.

Most of them see to have similar failures (result is already completed), so I wonder if there is an async task still running or a previous test that isn't cleaned up.

Assignee: nobody → jonalmeida942
Whiteboard: [geckoview:m97?] → [geckoview:m98]

I focused on bug 1642673, but I'm certain they are all related.

I thought this was an async bug, so I tried to see if using the looper from the handler we created for the GeckoResult. That still failed intermittently so I guess it seems like that's not on the same looper as the test is running on GeckoSessionTestRule. We do still want to complete on the main looper though, so maybe this a wasted effort to verify that it's an async bug.

Another long shot was to try post the completed message immediately instead of a 100ms delay - none of the other tests need this and it wasn't obvious to me why we're doing this here (maybe this was to overcome the flakiest originally and now it's showing up?).

Severity: -- → S3
Priority: P2 → P1
Whiteboard: [geckoview:m98] → [geckoview:m98][geckoview:m99]
Whiteboard: [geckoview:m98][geckoview:m99] → [geckoview:m98][geckoview:m99][geckoview:m100]

Unassigned for now since I'm not actively working on it.

Assignee: jonalmeida942 → nobody
Assignee: nobody → bugzeeeeee
Whiteboard: [geckoview:m98][geckoview:m99][geckoview:m100] → [geckoview:m98][geckoview:m99][geckoview:m100] [geckoview:m101]
See Also: → 1765421
Whiteboard: [geckoview:m98][geckoview:m99][geckoview:m100] [geckoview:m101] → [geckoview:m98] [geckoview:m99] [geckoview:m100] [geckoview:m101] [geckoview:m102]

Won't be fixed in 102, so carrying forward to 103.

Whiteboard: [geckoview:m98] [geckoview:m99] [geckoview:m100] [geckoview:m101] [geckoview:m102] → [geckoview:m98] [geckoview:m99] [geckoview:m100] [geckoview:m101] [geckoview:m102] [geckoview:m103]

Before this patch, in OnSearchResult, we would first call ProcessResult, and then sometimes PostSearchCleanup, and these two functions have duplicate logic. This patch reconciles the both in ProcessResult.

Attachment #9282843 - Attachment description: Bug 1738293 - Clean up the ProcessResult logic to avoid doing the same things twice → Bug 1738293 - Clean up the AutocompleteController logic to avoid doing the same things twice

Irene will land the fix after the long weekend.

Whiteboard: [geckoview:m98] [geckoview:m99] [geckoview:m100] [geckoview:m101] [geckoview:m102] [geckoview:m103] → [geckoview:m104]
Pushed by istorozhko@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/e070e403b23d
Clean up the AutocompleteController logic to avoid doing the same things twice r=dimi,sgalich
Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → FIXED
Target Milestone: --- → 104 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: