Bug 30811 – REGRESSION(49818): Tiger fails new fast/js/navigator-language.html test

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

http://build.webkit.org/results/Tiger%20Intel%20Release/r50133%20(5583)/results.html
http://trac.webkit.org/browser/trunk/LayoutTests/fast/js/navigator-language.html

Tiger returns "en" instead of "en-us" as expected.  We could fix the Tiger codepath, or easier would be to just update the Tiger results.

[Eric Seidel (no email)]

http://trac.webkit.org/changeset/49818

[Eric Seidel (no email)]

Here is the file for the mac:
http://trac.webkit.org/browser/trunk/WebCore/platform/mac/Language.mm

The implementation is in WebKit:
http://trac.webkit.org/browser/trunk/WebKit/mac/WebCoreSupport/WebViewFactory.mm#L129

Which calls through to here:
http://trac.webkit.org/browser/trunk/WebKit/mac/Misc/WebNSUserDefaultsExtras.m

I'm not sure what part of that ends up different on Tiger though.  None of it is special-cased for Tiger.

It's possible that "WKCopyCFLocalizationPreferredName" has a special Tiger implementation and thus is the root of the difference here.

Again, for now I think we should just check in the "failing" result for Tiger.

[Eric Seidel (no email)]

Really someone at Apple should confirm that we don't care about Tiger being different here.  I expect we don't.

[Chang Shu]

(In reply to comment #3)
> Really someone at Apple should confirm that we don't care about Tiger being
> different here.  I expect we don't.

Eric, I can add the special expected result for Tiger.

[Chang Shu]

Created attachment 42025 [details]
fix regression

[WebKit Commit Bot]

Comment on attachment 42025 [details]
fix regression

Rejecting patch 42025 from commit-queue.

[e-mail] does not have committer permissions according to http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/modules/committers.py.

[Eric Seidel (no email)]

Comment on attachment 42025 [details]
fix regression

This result looks wrong?  I don't see how Tiger would produce this result.  It's going to say FAIL instead of PASS.  You should pull the result off the tiger bot instead of creating it by hand.
http://build.webkit.org/results/Tiger%20Intel%20Release/r50217%20(5641)/fast/js/navigator-language-actual.txt

[Chang Shu]

(In reply to comment #7)
> (From update of attachment 42025 [details])
> This result looks wrong?  I don't see how Tiger would produce this result. 
> It's going to say FAIL instead of PASS.  You should pull the result off the
> tiger bot instead of creating it by hand.
> http://build.webkit.org/results/Tiger%20Intel%20Release/r50217%20(5641)/fast/js/navigator-language-actual.txt

You are right, Eric. Thanks!

[Chang Shu]

Created attachment 42040 [details]
fix result content

[Eric Seidel (no email)]

Comment on attachment 42040 [details]
fix result content

LGTM.  The commit-queue will close the bug.  I'm not sure if we need a bug open about tiger's misbehavior here or not.  I guess I'll assume not for now.  If one of the apple folks feels otherwise they should feel free to re-open.

[Eric Seidel (no email)]

Thank you very much for taking care of this Chang.

[WebKit Commit Bot]

Comment on attachment 42040 [details]
fix result content

Clearing flags on attachment: 42040

Committed r50223: <http://trac.webkit.org/changeset/50223>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Chang Shu]

(In reply to comment #11)
> Thank you very much for taking care of this Chang.

My pleasure. I usually panic when I see my patches cause regression. :)