Here's a quote from an e-mail I have from a while ago when I asked George Hoffman that question: > Well, in a drawing model that was a little better defined, in terms > of > hit-rules, etc. these could be done in a number of ways... generating > paths from the primitives and then using path-based clipping, for > example. > > The BeOS drawing model was never unified enough to be able to do > something like this, so ClipToPicture instead draws into a special > "accelerated" canvas that takes the pixel spans that are generated > and > converts them into BRegion rectangles and merges them with everything > that has been drawn thus far. That BRegion (or its inverse) is then > used to clip the drawing. In a lot of ways our server is a completely different animal from R5's, but at first glance I think that this doesn't sound too bad for us. Thoughts, guys? --DW