Re: [Gegl-developer] partial implementation of the change to gegl_scanline_processor

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sat, Mar 15, 2003 at 06:08:57PM -0800, Daniel Rogers wrote:
> Calvin Williamson wrote:
> > I think this looks more or less like the right idea. If you 
> > want to finish off changing all the scanline functions to 
> > just take the data lists, you could have a go at that, and
> > then we can commit after that happens. If not, I might
> > be able to do it in a few days...
> > 
> 
> Yeah, I was starting to head in that direction.
> 
> I have run into a bit of a problem though.  Validate inputs doesn't
> produce a GList of data_inputs that includes all the non-image inputs
> (for example, in GeglColor, there is a pixel, width, and height input
> that isn't registered as data_inputs.  I do see that these options are
> supposed to be stored in an image op itself.  I thought that a goal of
> your design was to allow this stuff to be registered as an input from
> another node?  If so, they shouldn't be stored in the op itself should they?
> 
> --
> Dan
> 

They are registered as inputs from another node as well. They
are also properties of this node. 

This is the problem I was talking about working on before. How
does validate_inputs take care of putting the appropriate 
fields from the op into each GeglData's value 
field before the data_inputs is passed back to the op. 

The issues that need to be resolved here are 
how attributes (properties) for an op can be both
settable during g_object_new and how they can be 
passed into the op as outputs from some other node. 

Heres an example of what I mean: 

After the registering the x,y as Scalar data and
the pixel as Pixel data, the GeglColor now officially
has 3 inputs. (Hence the badly named GeglNoInput really
means "no image input" in this case). You could 
hook up an op that outputs a scalar to either of the
x or y input slots and one that produces Pixel data
to the pixel slot. 

If those input nodes are not NULL, then the op
needs to use the "default" x,y,pixel 
that were passed in as regular properties to 
the op from g_object_new to put in the 
GeglData lists.

For now, GeglColor should just these values
directly from the fields in the op just like
it is currently doing.

Calvin   

[Index of Archives]     [Yosemite News]     [Yosemite Photos]     [gtk]     [GIMP Users]     [KDE]     [Gimp's Home]     [Gimp on Windows]     [Steve's Art]

  Powered by Linux