> ok, I was thinking the process worked differently. So the original > paper suggested that all those rects be stored in an image, but that > obviously breaks down if you want to have multiple threads working on > different parts of the image. Though if you have multiple threads > working on the same image, doesn't that mean they all have to have a > different GeglData? How is that implemented? I dont think we can store the data lists in GeglOp exactly like we are right now when we do multi-threading. They might go somewhere like in the EvalVisitor associated with the thread. Or you could store them on the Op but you have each thread look up its versions of the GeglData (which holds the extrinsic state for the op). I like storing it the Visitor, since each thread has its own version of this. > > I am not sure if this is correct in your view, but I see image > operations passing around objects that behave as images (e.g. the > buffers that image operations operate on are contained in actual > GeglImage objects). > I think that is my view too. Calvin
