Re: Last call for missing accessors

[Matthias Clasen <matthias clasen gmail com>]

On Thu, Feb 11, 2010 at 8:02 AM, Michael Natterer <mitch gimp org> wrote:
> On Mon, 2010-02-08 at 18:17 -0500, Matthias Clasen wrote:

> I am aware that this had to happen at some point :)

[...]

> So, we *will* ship GTK+ 2.20 with accessors missing. Especially
> accessors needed for subclassing, which access semi-internal
> stuff.
>
> Therefore I propose an exception to the "no new API in stable"
> rule so we can fulfill our promise of "port to GTK+ 2.20, and simply
> recompile against GTK+ 3.0". That exception would of course only
> apply to accessors for sealed members and to *nothing*else*.
>
> I fear that if we don't allow these exceptions in GTK+ 2.20.x,
> then the probability of stuff being ported to GTK+ 2.90/3.0
> will be lower, thus jeopardizing the entire effort.


Yes, this will be unavoidable, unfortunately. We will continue to add
accessors in 2.20.x. Can we do this in way that avoids constantly
breaking apps by adding new deprecations ?

Now, this still leaves us with the question of how to reach a
plausible GTK3 release in the fall. I personally want to achieve a few
things:

- Merge the xi2 work of Carlos

- Finish the gdbus work of David

- If there is time left, look at the RI stuff