Re: [PATCH 1/2] stroke: Fix large line widths for fallback stroke shaper
Chris Wilson <[email protected]>
| Newsgroups | gmane.comp.lib.cairo |
|---|---|
| Message-ID | <[email protected]> |
On Tue, Mar 12, 2013 at 03:16:01PM -0700, Martin Robinson wrote: > Fix the test case line-width-tolerance for the fallback stroke shaper. > Instead of drawing quads between spline points, draw triangle based on > the actual spline edges. This roughly follows the approach of the > tristrip and polygonal shapers. Ok, the only reason I left this path unscathed was so that I had something to fallback to... However the change in spline handling does improve several test and perf results. Let's look for opportunities to refactor the common code as we do so. (Can be done afterwards, so that we have correct code first, then correct smaller code.) > This commit also replaces the fixed point > _lines_intersect_or_are_coincident with slower, floating point > intersection code. _lines_intersect_or_are_coincident was faultily > trying to skip doing an actual division. Hmm, can you do this as a separate patch. Preferably with another to try and fix the current code before replacing it. As you should be able to avoid the division to check the intersection against [0,1] so I want a closer look at what you consider the bug to be. -Chris -- Chris Wilson, Intel Open Source Technology Centre -- cairo mailing list [email protected] http://lists.cairographics.org/mailman/listinfo/cairo