These pixel snapping convenience functions were failing to take the Box's location into account, which could result in reporting bad values. Thanks to Julien for catching this.
Created attachment 136547 [details] Patch
Comment on attachment 136547 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=136547&action=review > Source/WebCore/ChangeLog:13 > + No new tests. Untestable until we switch to FractionalLayoutUnits. Did this show a difference once we switch to FractionalLayoutUnit? If not, it really sounds like we should try to land a test that would show the difference.
(In reply to comment #2) > (From update of attachment 136547 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=136547&action=review > > > Source/WebCore/ChangeLog:13 > > + No new tests. Untestable until we switch to FractionalLayoutUnits. > > Did this show a difference once we switch to FractionalLayoutUnit? If not, it really sounds like we should try to land a test that would show the difference. This is easy to test for once we've cut over to FractionalLayoutUnits. Currently though, the parameter that was changed is unused :p I've uploaded a test case to our sub-pixel branch that will be included when we "flip the switch." You can see it here: http://trac.webkit.org/browser/branches/subpixellayout/LayoutTests/fast/sub-pixel/client-width-height-snapping.html
Comment on attachment 136547 [details] Patch OK.
Created attachment 136787 [details] Patch for landing
(In reply to comment #5) > Created an attachment (id=136787) [details] > Patch for landing Added a comment in the changelog explicitly stating that the test is forthcoming.
Comment on attachment 136787 [details] Patch for landing Clearing flags on attachment: 136787 Committed r113932: <http://trac.webkit.org/changeset/113932>
All reviewed patches have been landed. Closing bug.