Hi,

I have an application pane with :display-time :command-loop, inside a (scrolling ...) form. For the most part, in this application, it is likely that the contents of this pane after a redisplay are similar to those before: so, I would like to be able to preserve the viewport position across redisplays.

I've come up with the following method on redisplay-frame-pane:

(defmethod redisplay-frame-pane :around ((frame gsharp) (pane gsharp-pane) &key force-p) (declare (ignore force-p)) (let ((transformation (sheet-transformation pane))) (multiple-value-bind (x y) (window-viewport-position pane) (multiple-value-bind (x1 y1) (untransform-position transformation x y) (call-next-method) (setf (window-viewport-position pane) (values x1 y1))))))

which meets the spec, I suppose -- I don't know how well-defined it is, but it does what I wanted it to in McCLIM. However, it leads to visual artifacts: the call-next-method here eventually calls (window-clear), which resets the space-requirements and the transformation of the sheet, which means the scrollbars change size; then the whole of the redraw takes place, and only then does the viewport's position get reset.

My question then is twofold. Firstly, is this a sensible thing to be doing at all, or is there some other mechanism for achieving what I want here? (Incremental redisplay might be it, but that has some other problems in this application). Secondly, is there a way of eliminating these glitches?

Cheers,

Christophe