I'm pretty sure that NVDA sometimes makes use of 
IAccessible::text(0,IA2_TEXT_OFFSET_LENGTH,text)

Where IA2_TEXT_OFFSET_LENGTH is -1.

I'm not sure when this was introduced to IA2, but Gecko supports it, not 
sure about Symphony.

Very useful as it stops the need for calling nCharacters.

Mick

On 16/04/2010 7:40 AM, James Teh wrote:
> On 16/04/2010 2:39 AM, Carolyn MacLeod wrote:
>> Do AT's use IA2_TEXT_BOUNDARY_ALL?
> We don't. However, see below.
>
>> Or do they use IAccessibleText::text(0, length, pbstrText) to retrieve
>> all of the text?
> We do this.
>
>> 2) Technically, we don't need it because IAccessibleText::text(0,
>> length, pbstrText) already returns the whole text.
> Note that this requires ATs to make an extra call
> (IAccessibleText::nCharacters) to determine the length. Extra calls do
> matter across process boundaries. We do this anyway, but it's worth
> adding to the discussion.
>
>> I am tempted to return E_INVALIDARG if get boundaryType ==
>> IA2_TEXT_BOUNDARY_ALL in textAfterOffset and textBeforeOffset. Would
>> that break anybody?
> Hmm. That makes it look like the caller is wrong to try to use it when
> it is in the spec. I'd prefer E_NOTIMPL, but I realise that's only
> supposed to be used when the entire method isn't implemented. Are there
> any other options that indicate that this particular request isn't
> implemented?
>
> Jamie
>

-- 
Michael Curran
email/msn/jabber: [email protected]
Skype/Twitter: md_curran
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/accessibility-ia2

Reply via email to