oj@apollo.UUCP (11/30/87)
The rectangles delivered via expose events to an X11 client tesselate the exposed region of the window into horizontal strips. This appears to be a detail of the implementation of the sample server, in that the protocol doesn't specify how the tesselation should be carried out. I have some questions about this: (1) does it make sense in light of the protocol for clients to respond to series of expose events by accumulating the rectangles, then sending them to the server as a clip-rectangle list, then redrawing? (2) Would the present sample server handle this reasonably efficiently? (3) If the question to (1) is yes...Then it might be nice to specify an meaningful ordering parameter to the SetClipRectangles request when sending down the rectangle list. With the present sample server, this ordering would be YXSorted (right?). But the YXSorted nature of the Expose rectangles is an implementation detail, not architectural. Should the protocol spec be revised to specify that expose rectangles be ordered in a certain way? Perhaps a new code value for the ordering parameter, meaning "Expose-event order" ? /oj