Bug 120670 – REGRESSION(r154976): 7 accessibility tests failing on GTK

• Component: WebKitGTK
• Status: RESOLVED FIXED

[Zan Dobersek]

The following seven accessibility layout tests are failing since r154976:
http://trac.webkit.org/changeset/154976

accessibility/lists.html
accessibility/plugin.html
accessibility/table-detection.html
accessibility/table-one-cell.html
accessibility/table-with-aria-role.html
accessibility/table-with-rules.html
accessibility/transformed-element.html

http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=accessibility%2Flists.html%2Caccessibility%2Fplugin.html%2Caccessibility%2Ftable-detection.html%2Caccessibility%2Ftable-one-cell.html%2Caccessibility%2Ftable-with-aria-role.html%2Caccessibility%2Ftable-with-rules.html%2Caccessibility%2Ftransformed-element.html

All the failures are adding the aria-sort attribute to AXPlatformAttributes:
-AXPlatformAttributes: toolkit:WebKitGtk
+AXPlatformAttributes: aria-sort:AXUnknownSortDirection, toolkit:WebKitGtk

A couple of test baselines were updated in the original revision to include the new attribute. Is it also OK to rebaseline these 7 tests?

[Mario Sanchez Prada]

(In reply to comment #0)
> [...]
> A couple of test baselines were updated in the original revision to include the new attribute. Is it also OK to rebaseline these 7 tests?

If a new attribute has been supported and therefore exposed as an AtkAttribute is normal it shows up in the output for some tests now, so I'd say it's perfectly OK to rebaseline them all.

[Mario Sanchez Prada]

Zan already rebased expectations in r155033 and I just removed them from the TestExpectations file in r155446, so this is no longer an issue.