762798 - l10n tests shouldn't make use of window title but windowtype property

Status: RESOLVED FIXED

(Mozilla QA Graveyard :: Mozmill Tests, defect)

Description

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

I'm absolutely not sure what's going on here but some locales are failing by checking a sub window of the preferences dialog. One example is 'is':

http://mozmill-ci.blargon7.com/#/l10n/report/fdec829b93b19c73985be1d38878292c

Not sure yet, which window it is. I will have to find this out next.

Comment 1

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

So in the 'is' case it's because of a broken property string:

http://hg.mozilla.org/releases/l10n/mozilla-aurora/is/file/195eb83e843e/browser/chrome/browser/preferences/preferences.properties#l19

cookiepermissionstitle=Undanþágur - Smákökur\n

The final backslash shouldn't be there. That means we should always try to get the window by type and not by title. So we wouldn't be affected by those issues.

Comment 2

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

Here an example with the fix in-place:
http://mozmill-crowd.blargon7.com/#/l10n/report/87961186c2b807b7747aa7e6d4076279

Comment 3

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

Attachment: Patch v1

Comment 4

[Axel Hecht [:Pike]]

Comment on attachment 631323 [details] [diff] [review]
Patch v1

Review of attachment 631323 [details] [diff] [review]:
-----------------------------------------------------------------

I guess I can do a feedback+ on this patch, but not a real review, in particular as I have no estimate on teh impact of the dom-utils changes.

The general idea to identify windows by type sounds good to me, too.

I wonder if the "surprising" titles would actually make a good test themselves?

Comment 5

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

Comment on attachment 631323 [details] [diff] [review]
Patch v1

(In reply to Axel Hecht [:Pike] from comment #4)
> I guess I can do a feedback+ on this patch, but not a real review, in
> particular as I have no estimate on teh impact of the dom-utils changes.

Alright. The changes there only affect l10n tests and no other test-run.

> The general idea to identify windows by type sounds good to me, too.
> 
> I wonder if the "surprising" titles would actually make a good test
> themselves?

Yes, all that should be a separate test where we check the translation and functionality of items. But that's a whole different story we will not have time to work on in the foreseeable future.

Anthony, mind giving a review today? This failure kills some of our testruns and I would like to get it checked into each branch. Thanks.

Comment 6

[u279076]

Comment on attachment 631323 [details] [diff] [review]
Patch v1

Review of attachment 631323 [details] [diff] [review]:
-----------------------------------------------------------------

Patch looks fine to me. Please land.

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 07/21 - 08/08)]

Pushed:
http://hg.mozilla.org/qa/mozmill-tests/rev/9d3e8ba3cc0a (default)
http://hg.mozilla.org/qa/mozmill-tests/rev/2cfbb59eda4e (aurora)
http://hg.mozilla.org/qa/mozmill-tests/rev/05ebf59621ad (beta)
http://hg.mozilla.org/qa/mozmill-tests/rev/16249c8eb03b (release)
http://hg.mozilla.org/qa/mozmill-tests/rev/7f4451a79c25 (esr10)