The Chromium port is experimenting with running Skia as the WebKit rendering engine, and CoreGraphics as the UI rendering engine. This permits Chromium to unify its graphics story while leveraging OS X to draw elements like scrollbars and buttons. Restructure the common graphics units, points and rectangles, to convert to Sk-types and CG-types at the same time. A CL describing this change in more detail is forthcoming.
Created attachment 88408 [details] patch to change preprocessor conditions
Comment on attachment 88408 [details] patch to change preprocessor conditions View in context: https://bugs.webkit.org/attachment.cgi?id=88408&action=review > WebCore/ChangeLog:12 > + https://bugs.webkit.org/show_bug.cgi?id=57848 Note that typically this is: bug title bug link description if any Per file comments (which may be ditto). But ok this time.
Are these the only |#if PLATFORM(CG)| changes you will need? Would it be worth to add USE(CG_PRIMITIVES) and define that if PLATFORM(CG) || (USE(SKIA)…), for less repetiton?
Nico: There are other places where CG-specific code is required. However, these files together make up a nice chunk and are interdependent. I thought about creating a new preprocessor definition, but did not see precedent for doing things that way. The current way, while verbose, is completely self-documenting, where USE(CG_PRIMITIVES) would hide slightly that this code is used when CoreGraphics is the platform graphics implementation, and when Skia+Chromium+Darwin is chosen. However, I'm happy to change it to match WebKit conventions.
I guess if levin is fine with it, then the current patch matches WebKit conventions; the other way just seemed cleaner to me personally.
Created attachment 89438 [details] Patch
Comment on attachment 89438 [details] Patch LGTM.
Comment on attachment 89438 [details] Patch Clearing flags on attachment: 89438 Committed r83818: <http://trac.webkit.org/changeset/83818>
All reviewed patches have been landed. Closing bug.