Re: Text Attribute API revisions

[Brian Cameron <Brian Cameron Sun COM>]

Bill:

> I agree, and think you're right about "ref_attribute_bounds" except that the 
> semantics are confusing then (sounds like the "attribute bounds" are getting 
> referenced) - perhaps call it
> 
> atk_text_ref_run_attributes ()
> 
> [meaning, return a reference to the attributes over a run (the run bounds 
> being returned in the two "out" parameters) ]

Sounds good.  By the way, I'm now thinking that glib keyed data lists
aren't so useful.  the "foreach" function wants you to pass in a function
it calls for each element.  This seems cumbersome.  

The glib hash and dataset suffer from the same problem.  I think it might
be necessary for us to define our own AtkAttributeSet object.

Ah well.

Brian