Closed Bug 1599346 Opened 5 years ago Closed 4 years ago

Intermittent bugs/315920-9.html == bugs/315920-9-ref.html | image comparison, max difference: X, number of differing pixels: Y

Categories

(Core :: Layout, defect, P5)

defect

Tracking

()

RESOLVED FIXED
92 Branch
Tracking Status
firefox92 --- fixed

People

(Reporter: intermittent-bug-filer, Assigned: dholbert)

Details

(Keywords: intermittent-failure)

Attachments

(1 file)

The reftest screenshot shows that the outline of an un-ticked <input type="radio"> element rendered a bit differently between testcase & reference case. (It might be just shifted slightly, by a sub-pixel amount?)

Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → INCOMPLETE
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---
Summary: Intermittent bugs/315920-9.html == bugs/315920-9-ref.html | image comparison, max difference: 37, number of differing pixels: 65 → Intermittent bugs/315920-9.html == bugs/315920-9-ref.html | image comparison, max difference: X, number of differing pixels: Y
Status: REOPENED → RESOLVED
Closed: 5 years ago5 years ago
Resolution: --- → INCOMPLETE
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---
Status: REOPENED → RESOLVED
Closed: 5 years ago5 years ago
Resolution: --- → INCOMPLETE
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---

The failure in comment 0 had max difference: 37, number of differing pixels: 65 (and looked like maybe the radio button was being rendered at a slightly different subpixel position, though not in a clearly broken way aside from the nondeterminism)

The failure in comments 5 and 9 can't be viewed anymore, unfortunately (the parsed log has expired and we don't have the raw log link here).

The failure in comment 12 has max difference: 1, number of differing pixels: 5 which is much more clearly un-interesting -- just some as-subtle-as-possible antialiasing differences around a few pixels at the edge of the radio buttons' rings here.
(raw log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/d4dP-imeSFqmKoGELMtB7g/runs/0/artifacts/public/logs/live_backing.log )

Anyway: if this happens again, we should probably just add a fuzzy annotation that covers the recent observed fuzzy-mismatch levels. At this point, none of the failures here seem particularly alarming or merit spending too much time investigating.

Status: REOPENED → RESOLVED
Closed: 5 years ago4 years ago
Resolution: --- → INCOMPLETE

For the record: the comment 16 failure is max difference: 1, number of differing pixels: 41 (with the differing pixels being antialiased fringe at the curved edge of the radio-button's circle).

Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---

comment 18 is the same sort of failure as comment 16 (with max difference: 1, number of differing pixels: 41, on linux1804-64-asan opt).

This is to allow "max difference: 1, number of differing pixels: 41". The
differing pixels are all antialiased fringe along the curved edges of the
test's radio-buttons.

Assignee: nobody → dholbert
Pushed by dholbert@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/be5486992cee Annotate recent observed fuzziness in reftest 315920-9.html. r=intermittent-reviewers,ahal
Status: REOPENED → RESOLVED
Closed: 4 years ago4 years ago
Resolution: --- → FIXED
Target Milestone: --- → 92 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: