949428 - Test failure "Search bar has the focus - '[object HTMLInputElement]' should equal '[object XULElement]'" in testSearchViaShortcut.js

Status: RESOLVED FIXED

(Mozilla QA Graveyard :: Mozmill Tests, defect)

Description

[Andreea Matei [:AndreeaMatei]]

Started failing today with Nightly, so far on all OS X machines. Might be a regression.

Mario, could you please check this out? And please prepare a skip patch if it turns out it's reproducible all the time.

Comment 1

[mario garbi]

Attachment: skip.patch

Added a skip patch for OSX platforms only as I cannot reproduce it on Linux or Windows locally and we only had OSX failures on CI.
Coming with a fix patch as soon as possible.

Comment 2

[Andreea Matei [:AndreeaMatei]]

Comment on attachment 8346526 [details] [diff] [review]
skip.patch

Review of attachment 8346526 [details] [diff] [review]:
-----------------------------------------------------------------

Disabled:
http://hg.mozilla.org/qa/mozmill-tests/rev/bb4513f7af0b (default)

Comment 3

[Andreea Matei [:AndreeaMatei]]

This fails on Linux as well. Please do a follow up to skip it completely.
http://mozmill-daily.blargon7.com/#/functional/report/ca1869364f98dd0de71202e01e14f7ad

Comment 4

[mario garbi]

Attachment: Patch_v2

Followup skip patch for all platforms.

Comment 5

[Andreea Matei [:AndreeaMatei]]

Comment on attachment 8346553 [details] [diff] [review]
Patch_v2

Review of attachment 8346553 [details] [diff] [review]:
-----------------------------------------------------------------

http://hg.mozilla.org/qa/mozmill-tests/rev/2ca8396f9bdf (default)

Comment 6

[Cosmin Malutan, [:cosmin-malutan]]

This fails due to focusing the search bar via shortcut, if we change to anything else it will pass.

This was introduced by bug 901207, changeset http://hg.mozilla.org/mozilla-central/rev/3cafeb339fac

The Cmd + k / Cmd + j (linux) should focus the search bar, but instead it navigates to "Search engine" home page. This happened by the changing of the BrowserSearch:webSearch method (see the changeset).

If this is an expected behavior we should change our test.

Jared can you please confirm that this is an expected behavior?

Comment 7

[Jared Wein [:jaws] (please needinfo? me)]

This was not an expected behavior but was just fixed in today's Nightly build (see https://bugzilla.mozilla.org/show_bug.cgi?id=949365).

You can either reenable your test, or remove the mozmill test now since we have a duplicate test in the mochitest-browser suite: http://mxr.mozilla.org/mozilla-central/source/browser/components/customizableui/test/browser_901207_searchbar_in_panel.js

Comment 8

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Thanks Jared for fixing this issue. In terms of removing, we don't want to do that given that this test is covering a keyboard shortcut which can be different in each locale. Mochitest only runs in en-US while we run nearly each locale. So keeping this test has a very high value. Lets reenable it.

Comment 9

[Jared Wein [:jaws] (please needinfo? me)]

Ok, thanks for the explanation. Glad that it was resolved so quickly.

Comment 10

[Andrei Eftimie]

Attachment: unskip.patch

Unskipped the test.
It passes with the latest Nightly.

Comment 11

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Comment on attachment 8347902 [details] [diff] [review]
unskip.patch

Review of attachment 8347902 [details] [diff] [review]:
-----------------------------------------------------------------

Looks fine. Please land it as soon as you can so we can verify the fix across all platforms.

Comment 12

[Andrei Eftimie]

Unskip landed:
http://hg.mozilla.org/qa/mozmill-tests/rev/400a2a2447ba (default)

We should be good here.