1592041 - Debian 10 - layout/reftests/bidi/reftest.list | multiple failures in brackets tests

Status: RESOLVED INVALID

(Core :: Layout: Text and Fonts, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: egao [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=271937786&repo=try
Full log: https://queue.taskcluster.net/v1/task/Vhp9Ri3XRAu_wvsK4MvvLw/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://queue.taskcluster.net/v1/task/Vhp9Ri3XRAu_wvsK4MvvLw/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

---

suite: reftest-4

context:

All tests in the manifest that are highlighted fail with similar differences: https://searchfox.org/mozilla-central/source/layout/reftests/bidi/reftest.list#163-172

Sometimes the failure is difference: 1 pixels: 1.
Other times the difference is larger at 64.

[task 2019-10-18T17:33:52.685Z] 17:33:52 INFO - REFTEST TEST-START | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr-ref.html
[task 2019-10-18T17:33:52.685Z] 17:33:52 INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr.html | 158 / 177 (89%)
[task 2019-10-18T17:33:52.724Z] 17:33:52 INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr-ref.html | 158 / 177 (89%)
[task 2019-10-18T17:33:52.948Z] 17:33:52 INFO - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-ltr-ref.html | image comparison, max difference: 1, number of differing pixels: 3

[task 2019-10-18T17:33:52.957Z] 17:33:52 INFO - REFTEST TEST-START | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl-ref.html
[task 2019-10-18T17:33:52.957Z] 17:33:52 INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl.html | 159 / 177 (89%)
[task 2019-10-18T17:33:52.958Z] 17:33:52 INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl-ref.html | 159 / 177 (89%)
[task 2019-10-18T17:33:53.118Z] 17:33:53 INFO - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/brackets-2a-rtl-ref.html | image comparison, max difference: 64, number of differing pixels: 140

Comment 1

[Daniel Holbert [:dholbert]]

Looks like the failures in these "brackets" test are nearly all on ligatures.

The test with max difference: 64, brackets-2a-rtl.html, also has its large "b" character at a slightly different position (off by ~1/2px), which makes all of its edge pixels have a substantially different shade. This is a bit odd since it's a monospace font... I'm guessing this is a quirk of the particular platform font that's being used.

Anyway, all of these brackets tests (including this one with the the subtly-mismatching "b") seem suitable for fuzzy annotations.

Comment 2

[Daniel Holbert [:dholbert]]

(In reply to Daniel Holbert [:dholbert] from comment #1)

Anyway, all of these brackets tests (including this one with the the subtly-mismatching "b") seem suitable for fuzzy annotations.

This goes for all of the tests in this test-run, except for the first one, 1082844.html -- that one has some pieces of characters entirely clipped for some reason, and that seems maybe like a real bug (if it's reliable) or perhaps a reason to adjust the test itself (if there's a reasonable explanation for it.)

Comment 3

[Edwin Takahashi (:egao | infrequent contributor)]

Closing this bug due to redirection of efforts from debian10 to ubunu1804 image, and this bug is not reproducible on the ubuntu1804 image.