Re: Problem with __plt_get_axis_arg__.m

[Michael Goffioul]

On 12/3/07, David Bateman <address@hidden> wrote:
> Ouch.. Ok I think I see what you are getting at, but its my belief that
> __patch__ should expect only an axes handles, and the the hggroup should
> be handled in patch.. Looking at
>
> http://www.mathworks.com/access/helpdesk/help/techdoc/index.html?/access/helpdesk/help/techdoc/ref/patch.html
>
> there isn't supposed to be a manner to call patch with an hggroup or in
> fact an axes handle..

Indeed...

I think I was misled by this first argument to the patch function, assuming
it should be the way to specify the patch's parent. Actually, I think the
parent, if specified, should be given through the property/value pairs and
not as first argument of patch.m. This seems to be valid for all core
objects (line, patch, image, light, text, hggroup and hgtransform).

Then I know I'll face another problem in JHandles (which I was trying
to handle with this first argument of patch). In my current implementation,
the parent of an object should be known at creation-time: reparenting is
not really supported yet. In the graphics code, the current pattern is
to create an empty object (e.g. __go_patch__()) and then fill its properties.
To have more flexibility (and also avoid current reparenting issues),
I'd like to have the object creation function called with all init properties,
instead of the current 2-step process. Is this feasible?

Note: this also would provide more efficient JHandle code, because
listeners are not active at creation-time. So doing

    h = __ go_patch__ ();
    set (h, "facevertex", fv);

is slower than doing

    h = __go_patch__ ("facevertex", fv);

Michael.