Re: [g-a-devel]Implementing support for state sets in at-spi

[Michael Meeks <michael ximian com>]

Hi Bill,

On Mon, 2002-02-11 at 15:01, Bill Haneman wrote:

> The intention was for these operations to be opaque.  I agree that we
> need some kind of data marshalling but my intention was never to expose
> it; one reason has to do with language bindings, another was the desire
> not to expose implementation-dependent marshalling/demarshalling code.

	Fine.

> That said, the implementation code for compare and isEqual would like
> *some* way to do its work efficiently, without having to call 'contains'
> multiple times.  Do we have a convention for doing this, i.e. for
> "private" IDL methods?  That's what we'd want here I think.

	No - there is no convention.

> We could have Accessibility_StateSet implement some OpaqueAny thingy
> kind of interface which user code would expressly not be able to do
> anything useful with, but perhaps we can just add a getPackedData
> method to StateSet, with the proviso that user code shouldn't muck
> with it?

	Sounds like a good enough idea.

+	Any getPackedData ();

	Ultimately I don't believe that exposing the getData method as:

+	sequence<StateType> getData ()

	would hurt the API whatsoever, and would quite possibly help other
people make the optimizations necessary in some cases. I don't see how
that would make the API any less opaque than it is already.

	But as you like; either way we need the new method; can you decide what
you want, propose it to the release team, etc.

	Regards,

		Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot