The figure interface

A figure can be contained in one or more figure sets (see the next page for more on figure sets.) One of these must be the parent returned by getParent(). In the Diva canvas, FigureContainer is implemented by FigureLayer and AbstractFigureContainer, so the parent can either be a layer or another figure.

A figure has two methods that return geometry-related objects: getBounds() and getShape(). The shape is intended for use by anything that needs a reasonable estimate of the figure's actual shape, but figures that have very complex shapes may choose to simply return the bounding box. Both the bounds and the shape are given in the enclosing transform context of the figure.

Most of the remaining methods in the Figure interface deal with the figure geometry. The transform() and translate() methods change the figure's coordinates, either directly or by changing the figure's associated transform context; contains() and intersects() test if the figure contains or intersects the given rectangle; hit() is similar to intersects(), but takes account of transparent regions (such as the inside of an unfilled rectangle).

Note that Figure inherits repaint() from CanvasComponent. This method propagates a damage region up the tree, until it gets to the JCanvas, at which point the appropriate region of the JCanvas is marked damaged. (This is the region that will later be passed to paint() if there are no other intervening calls that also damage the canvas.) repaint() must be called by every client that changes the appearance of the figure, to ensure that the change becomes visible.