1124549 - reftests/canvas/784573-1.html fails on OS X 10.10

Status: RESOLVED FIXED

(Core :: Graphics: Canvas2D, defect)

Description

[Phil Ringnalda (:philor)]

Attachment: reftest.log

try: -b o -p macosx64 -u reftest[10.10] -t none

https://treeherder.mozilla.org/logviewer.html#?job_id=4381340&repo=try

16:30:23 INFO - REFTEST TEST-LOAD | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/canvas/784573-1.html | 4223 / 11728 (36%)
16:30:23 INFO - REFTEST TEST-LOAD | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/canvas/784573-1-ref.html | 4223 / 11728 (36%)
16:30:24 INFO - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/canvas/784573-1.html | image comparison (==), max difference: 1, number of differing pixels: 128
...
16:30:24 INFO - REFTEST INFO | Loading a blank page
16:30:24 INFO - REFTEST TEST-END | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/canvas/784573-1.html

Comment 1

[Phil Ringnalda (:philor)]

Oh, is this just the same thing that the existing AzureQuartz fuzz is for, only this brings it up to 128 instead of 113 pixels?

Comment 2

[Phil Ringnalda (:philor)]

Indeed, had I just looked at the annotation for this, and the annotation for the other things that were also tens of thousands of pixels off by one or two on AzureQuartz that were also just a few pixels short, I could have figured that out myself.

Adjusted in https://hg.mozilla.org/integration/mozilla-inbound/rev/a18024a0bd19

Comment 3

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/a18024a0bd19