Okay, time for an update. I've spent the last couple of weeks working on a scanbeam-based polygon clipping algorithm, and at around 11pm last night the pieces all fell into place nicely. I've got a simple, robust method for handling the nightmare cases that break the current implementation and am starting the final implementation today. This will include a lot of comments and a separate README on what's happening. I'm really happy with how things have turned out, as the last year (and it must be about my one-year anniversary) has taught me a lot about computational geometry. I'm much more confident that my solution is correct this time, mostly because I designed it rather than relying on other algorithms, and it will be *fast*. Now that we have an open-source GLU from SGI, the only differentiating factor will be performance, so I'm going to make sure that my implementation is at least faster than SGI's :-) I'm expecting to have some initial 3.3 code ready in a couple of days. -- Gareth _______________________________________________ Mesa-dev maillist - [EMAIL PROTECTED] http://lists.mesa3d.org/mailman/listinfo/mesa-dev