Re: [g-a-devel] RFC: AtkText simplification

[Joanmarie Diggs <jdiggs igalia com>]

On 05/05/2013 09:37 PM, Trevor Saunders wrote:

> I wonder if there is a reason you prefer this approach to the ia2 one in
> which after / before varients are kept,  but only the actual contents of
> the word / line / whatever are returned not the word / line and the
> whitespace before or after depending?

This approach seems simpler to me:
* One method rather than three
* A given block of text is neatly chopped up with no "scraps" left over,
  thus the user is always in a defined text boundary.

Is there some reason why you want to implement three methods instead of
one? And given that there are four text boundaries (char, word, line,
sentence) that means handling 12 cases instead of four.

On a related note, there were very few instances in Orca of the END
boundary. And now there are none. Orca master is strictly relying upon
the AT flavor when getting text and is only getting text for the START
variant. the ATK and AT-SPI2 text simplification can safely begin now.
<smiles>

--joanie