[Accessibility-ia2] accNavigate relation queries

[Aaron Leventhal]

Right, it would never QI to both -- that wouldn't make sense afaict.

- Aaron

William W. Smith wrote:
> That solves the concern I had: I'll need to do the QI's on the 
> IDispatch  in the order first IAccessible and then if that fails try 
> IEnumVARIANT.    The case I was considering was when the IDispatch  
> QI'ed to both IAccessible and IEnumVARIANT at the same time.
>
> Bill
>
> Aaron Leventhal wrote:
>> Here's the rule:
>> 1) You'll always get an IDispatch pointer
>> 2) If it QI's to IAccessible, it's a single target
>> 3) If it QI's to IEnumVARIANT there are multiple targets
>> 4) If it's null there are no targets
>> 5) Nothing else is possible.
>>
>> Is that what you want from it?
>>
>> - Aaron
>>
>> William W. Smith wrote:
>>> Hi,
>>>
>>> How will the AT know whether the return value from the accNavigate 
>>> is the actual target of the relation vs. a  parent of multiple targets?
>>>
>>> I think that if the return value from accNavigate could be either a 
>>> UI object and a indirect reference to the targets, sometimes both 
>>> could be QI'd to IEnumVariant to get the children of the UI element 
>>> in a manner that  would be equivalent to the multiple target 
>>> technique. If I understand  it correctly, this would mean that the 
>>> result from the accNavigate needs to always return a IDispatch 
>>> representing a single object or always something that could be QI'd 
>>> to IEnumVarariant or but not a combination of the two to prevent the 
>>> results from being ambiguous.
>>>
>>> Bill
>>>
>>> accessibility-ia2-request at lists.freestandards.org wrote:
>>>  
>>>>> Larry,
>>>>>
>>>>> I have them documented here:
>>>>> http://www.mozilla.org/access/windows/at-apis#knowndif
>>>>>
>>>>> They cannot be programatically discovered.
>>>>>
>>>>> We don't support more than 1 relation target yet, but we can do 
>>>>> that by
>>>>> returning an IUnknown which can be QI'd to an IEnumVARIANT. That's
>>>>> similar to how get_accSelection() already works.
>>>>>
>>>>> - Aaron
>>>>>           
>>>
>>> _______________________________________________
>>> Accessibility-ia2 mailing list
>>> Accessibility-ia2 at lists.freestandards.org
>>> http://lists.freestandards.org/mailman/listinfo/accessibility-ia2
>>>
>>>   
>>
>