1139895 - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/bugs/1062792-1.html | image comparison (==), max difference: 1, number of differing pixels: 5

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Runtime, defect)

Description

[:gerard-majax]

So we have this test failing on Mulet, and it's not something skipped on B2G/B2G Desktop

Comment 1

[:gerard-majax]

Attachment: test execution screenshot

Comment 2

[:gerard-majax]

Attachment: reference screenshot

Comment 3

[:gerard-majax]

Timothy, do you have any idea why the test report failure ? The screenshots extracted from the try log shows nothing :(

Comment 4

[:gerard-majax]

Attachment: try_ubuntu64_vm-mulet_test-reftest-3-bm68-tests1-linux64-build44.txt.gz

Comment 5

[Timothy Nikkel (:tnikkel)]

In the tree there is a file called layout/tools/reftest/reftest-analyzer.xhtml, it's used to process reftest logs and analyze the failures when you open it in a browser. It has a circle differences option which is good for cases like this where the images look the same to human eyes. In this case it looks like there are tiny differences at the ends of two scrollbars that overlap. I think we should just add fuzz to the reftest to make it pass.

Comment 6

[:gerard-majax]

Oh, right. We have it on treeherder even: http://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=http://ftp.mozilla.org/pub/mozilla.org/b2g/try-builds/[e-mail]-627be3f63a32/try-linux64-mulet/try_ubuntu64_vm-mulet_test-reftest-3-bm52-tests1-linux64-build26.txt.gz&only_show_unexpected=1

Comment 7

[:gerard-majax]

(In reply to Timothy Nikkel (:tn) from comment #5)
> In the tree there is a file called
> layout/tools/reftest/reftest-analyzer.xhtml, it's used to process reftest
> logs and analyze the failures when you open it in a browser. It has a circle
> differences option which is good for cases like this where the images look
> the same to human eyes. In this case it looks like there are tiny
> differences at the ends of two scrollbars that overlap. I think we should
> just add fuzz to the reftest to make it pass.

You are right.

Comment 8

[:gerard-majax]

Attachment: Capture du 2015-03-06 15:58:27.png

This exposes the differences. It's around the junctions of scrollbar, and I'll go with fuzz for this.

Comment 9

[:gerard-majax]

Attachment: Add fuzz for 1062792-1.html r=...

Comment 10

[:gerard-majax]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=d6a0430596e0

Comment 11

[:gerard-majax]

(In reply to Alexandre LISSY :gerard-majax from comment #10)
> https://treeherder.mozilla.org/#/jobs?repo=try&revision=d6a0430596e0

broken for a typo ..

Comment 12

[:gerard-majax]

Attachment: Add fuzz for 1062792-1.html r=...

Comment 13

[:gerard-majax]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=97360aeaf382

Comment 14

[:gerard-majax]

Comment on attachment 8574036 [details] [diff] [review]
Add fuzz for 1062792-1.html r=...

Not sure if you can review, in case you cannot, thanks for forwarding :)

Comment 15

[:gerard-majax]

Attachment: Add fuzz for 1062792-1.html r=tn

Comment 16

[:gerard-majax]

Attachment: Add fuzz for 1062792-1.html r=tn

Comment 17

[:gerard-majax]

Comment on attachment 8574262 [details] [diff] [review]
Add fuzz for 1062792-1.html r=tn

Carrying r+

Comment 18

[:gerard-majax]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=97360aeaf382&exclusion_profile=false

Comment 19

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/29eb3c2805dc

Comment 20

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/29eb3c2805dc