Flip ImageOrientation coordinate system from lefthanded to righthanded
Created attachment 170727 [details] Patch
This is the promised follow-up patch. As expected, the skia code becomes a bit shorter, the cg code a bit longer. I don't care too much if we check this in or not. WebKit's coordinate system is generally top-left / right-handed, so this would make the world a tiny bit more consistent I guess.
Created attachment 170728 [details] Patch
Created attachment 170729 [details] Patch
Comment on attachment 170729 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=170729&action=review > Source/WebCore/ChangeLog:8 > + This is a pure refactoring and has no observable effects. Ideally you woudl explain teh justification in the ChangeLOg. Something along the lines of "platform/graphics, noteably GraphicsContext is defined to use a right handed coord system, yet ImageOrientation was an outlier from platform/graphicsin that it used a left-handed coord system (like how CoreGraphics does). This patch fixes ImageOrientation to match the rest of platform/graphics." or some such.
Created attachment 170731 [details] Patch
Done.
Consistency is good. We really should document some of platform/graphics better some day. http://trac.webkit.org/wiki/LayoutUnit has been helpful to me so many times. I wish we had more documentation like it.
Comment on attachment 170731 [details] Patch I'm a fan of consistency. Thank you for following up on this.
Noel: Hopefully we're all on the same page now? I look forward to your comments and reviews when you return.
Comment on attachment 170731 [details] Patch Clearing flags on attachment: 170731 Committed r132555: <http://trac.webkit.org/changeset/132555>
All reviewed patches have been landed. Closing bug.
(In reply to comment #10) > Noel: Hopefully we're all on the same page now? I look forward to your comments and reviews when you return. Sorry for the delayed reply, this went under my radar since it was not present in the meta bug. Let's add it there for posterity.
(In reply to comment #10) I also considered that we should order the orientation values numerically in ImageOrientation.cpp. I filed bug 101596 about it.