Re: AT-SPI standard's management of text objects

[Samuel Thibault <samuel thibault ens-lyon org>]

Ksamak, on Wed 16 Nov 2016 09:32:20 +0100, wrote:
> On Tue, Nov 15, 2016 at 07:49:00PM +0100, Samuel Thibault wrote:
> > > I know these behaviours are probably correct according to AT-SPI
> > > standard,
> >
> > What makes you think that?
> Well, i think there might be a reason for such a behaviour. I can
> imagine the orca screen reader wanting to ignore an empty line for
> example.

I believe it uses other ways to detect that.

> > > but is it voluntary?
> > I don't think it is. Probably just nobody ever had the need for these to
> > be fixed.
> >
> > > Is it possible to change any of that in the future, without impairing
> > > the other applications relying on at-spi?
> >
> > My guess would be yes. I don't see how useful it'd be to get 0,0
> > coordinates while the caret is obviously not at 0,0.
> Well, I was thinking the design was not thought through for visual
> applications. There is indeed no thickness to report for a caret that is
> on a newline.

Well, the caret does have some thickness, at least 1 pixel.

Samuel