Re: ISSUE-671: Re: Mapping @aria-invalid: string versus token value

This is already how we do it in Firefox, so this is OK AFAICT.

Marco



On Tue, Jul 29, 2014 at 4:26 PM, Joseph Scheuhammer <clown@alum.mit.edu>
wrote:

> Raised ISSUE-671: UAIG should expose aria-invalid as token values (as
>> defined per spec), not strings.
>> https://www.w3.org/WAI/PF/Group/track/issues/671
>>
>
> Here is a way to do this.  Can the AAPI experts advise please, aka
> Cynthia, David, Joanie, and James?  Thanks.
>
> There are three rows in the mapping table.  One is for
> aria-invalid="true", "spelling", or "grammar", while another is for
> aria-invalid="false".  Those two rows can stay as they are.
>
> The third row is the problem, where aria-invalid="unknown value", meaning
> a value outside the enumeration given in the spec.  Here are the current
> mappings, and the suggested changes.
>
> MSAA+UIA Express, and UIA:
> - Expose <value> in IsDataValidForForm property.
> Change to:  Expose 'false' in IsDataValidForForm property.
>
> MSAA+IAccessible2:
> - Set state IA2_STATE_INVALID_ENTRY".
> - Expose <value> as a text attribute.
> The first bullet doesn't need to change.  Change the second bullet to:
>  Expose "invalid:true" as a text attribute.
>
> ATK/AT-SPI (similar to MSAA+IA2):
> - Set state STATE_INVALID_ENTRY
> - Expose <value> as a text attribute.
> Change the second bullet to:  Expose "invalid:true" as a text attribute.
>
> AXAPI:
> AXInvalid:  <value>
> Change to:  AXInvalid: "true"
>
> --
> ;;;;joseph.
>
> 'Array(16).join("wat" - 1) + " Batman!"'
>            - G. Bernhardt -
>
>
>