Re: Are Out-of-Tree Widgets Second-Class Citizens?

[Allin Cottrell <cottrell wfu edu>]

On Thu, 1 Oct 2009, Cody Russell wrote:

> On Mon, 2009-09-28 at 15:12 -0400, Morten Welinder wrote:
> > This is a consequence of the halfway G_SEALing that was done.
> > Insofar G_SEAL is a good idea, it should apply to GTK+ itself,
> > i.e., GtkLabel has no business messing with the internals of
> > GtkWidget, although obviously it should have access to its own
> > internals...
>
> This is part of the plan for 3.0.  I think the original idea was
> that 2.18 or 2.90 would GSEAL everything, and 3.0 would actually
> remove those struct members.  This obviously means that it won't
> have access to those members anymore and it will have to use
> either the public APIs or the members of private structs.

Hmm. I had written earlier,

> > I'm not sure if the solution Morten advocates -- namely,
> > applying the GSEAL principle internally -- is the best one,
> > although maybe it is.

to which Brian Tarricone replied,

> It's not even necessary.  Gtk can access struct members directly
> if desired (for performance reasons, or whatever), but the API
> should expose them for subclasser use as well.

These two responses (Cody's and Brian's) don't jive. Any more
clarification available? Thanks.

Allin Cottrell