Re: States in GailLabel

[Bill Haneman <bill haneman sun com>]

Brian Cameron wrote:
> 
> Bill:
> 
> > On reflection, since our labels implement AtkText, I think this
> > probably would be useful, since it would be better to know in advance
> > whether there are multiple lines rather than attempt multiline
> > navigation and get "fail" type of result.
> 
> So are we saying we want to implement this state as synamic for
> labels?

As I understand it, it would only be dynamic if/when the label is
replaced.  We already must listen for such programmatic label changes,
so I think the GailLabel state should reflect this.  However if
resizeable labels do auto-word wrapping of some sort then we might
reconsider... I didn't think this was a situation that arose in GTK+
(resizeable labels + wordwrap).
 
> > Brian, what is the behavior of the Gail AtkText implementors if one
> > attempts to navigate past the end of the text buffer?
> 
> I'm not sure what you mean by "navigating past the end".  Do you
> mean by moving the "virtual caret"?

After a moment's thought I don't think 'navigating past the end' is
possible via the atk_text_get_text_[at,after,before]_offset functions
unless an invalid offset is passed.  What is the result if the offset
is invalid?

I notice that the ATK API docs on developer.gnome.org are considerably
out of date, they do not include the "end_offset" out-parameters for
atk_text_get_text_*.

-Bill

> 
> Brian

-- 
--------------
Bill Haneman
Gnome Accessibility / Batik SVG Toolkit
Sun Microsystems Ireland