Re: Expectations regarding aria-hidden="true" + a valid aria-label value?

We agreed that the presence of aria-label should cancel out role="none" or
role="presentation".

aria-hidden is different. You might have an element with a valid role and
label, which you are temporarily hiding (e.g., to reflect a change in
opacity or a movement off-screen).

The only thing that cancels out aria-hidden="true" is interactivity. If the
element receives keyboard focus, it can't be hidden.

~ABR

On 15 November 2016 at 10:22, Joanmarie Diggs <jdiggs@igalia.com> wrote:

> Hey all.
>
> I am looking into failures encountered when my ATTA is used with the SVG
> testable statements found here: [1].
>
> My recollection of the consensus reached was that the presence of a
> valid aria-label value necessitated inclusion in the accessibility tree
> -- even if the element with that aria-label also had aria-hidden="true"
> or was contained in an ancestor with aria-hidden="true". However, that
> seems to contradict the expectations in a significant number of tests on
> the aforementioned testable statements page.
>
> Could someone please clarify if my recollection is wrong or if the
> testable statements are? If it is the former, then I have an
> implementation to fix, but if it is the latter, it would be great to
> have the testable statements wiki corrected.
>
> Thanks!
> --joanie
>
> [1]
> https://www.w3.org/wiki/SVG_Accessibility/Testing/Test_
> Assertions_with_Tables_for_ATTA
>
>