Re: figure positioning problem

[Michael Goffioul]

On Sat, Nov 19, 2011 at 4:18 PM, John W. Eaton <address@hidden> wrote:
> Do you mean the following change?
>
> diff --git a/src/graphics.cc b/src/graphics.cc
> --- a/src/graphics.cc
> +++ b/src/graphics.cc
> @@ -3237,9 +3237,9 @@
>     {
>       Matrix old_bb, new_bb;
>
> -      old_bb = get_boundingbox ();
> +      old_bb = get_boundingbox (true);
>       position.set (v, true, do_notify_toolkit);
> -      new_bb = get_boundingbox ();
> +      new_bb = get_boundingbox (true);
>
>       if (old_bb != new_bb)
>         {

Yes, that's what I mean.

> If so, then yes, this avoids the problem for me.  But then the figure
> that I'm working with is also not resized in the way I think it is
> supposed to be.  So instead of the figure window being tight around
> the axes objects, there is a lot of whitespace padding.  But at least
> the axes are fully within the figure window.
>
> Should I make this change?

Yes. The current code is wrong: it'll use the external boundingbox,
which is not defined/set by the FLTK (or gnuplot) backend. So
basically, the test (old_bb != new_bb) will never be true (changing
the position property will not change the outerposition, so the
computed external bounding box will still be the same).

If it still doesn't do what you want, there might be another problem.
But at least the change above fixes the problem I've introduced when
adding the outerposition property.

Michael.