Steps: 1. Try to drag thumb to the half way point. 2. Notice it is not actually at the half way point.
Created attachment 27447 [details] Demo
Confirmed as a regression with r40771. The scroller is shorter than it should be, so it appears to be mispositioned when dragging.
<rdar://problem/6566875>
Adding Hyatt since he wrote http://trac.webkit.org/changeset/40771.
Oh, Alexey clarified that 40771 is just the revision he used to confirm the bug, not the cause of the bug.
This regressed between the r36847 and r36882 nightlies. It shouldn't be too hard to find the exact point of the regression, because it's such a small range.
This regressed in r36869: http://trac.webkit.org/changeset/36869
Created attachment 29704 [details] Proposed patch Do we usually add pixel tests for something like this?
Comment on attachment 29704 [details] Proposed patch r=me And, yes, a manual test or a pixel test or both would be good to add.
Assigning to cpst for test addition and landing.
cpst? Ping?
cpst: ping? Please advise.
Comment on attachment 29704 [details] Proposed patch r- for lack of tests per darin's comment above. Can't land this.
*** This bug has been marked as a duplicate of bug 34049 ***