All four of the hit-test-* layout tests fail. Could just be pixel differences, but might be actual bugs here.
I think this boils down to positionForPoint on RenderMultiColumnSet needing to work with logical coordinates instead of treating the point passed in as physical.
Created attachment 230193 [details]
Comment on attachment 230193 [details]
View in context: https://bugs.webkit.org/attachment.cgi?id=230193&action=review
> + log ("Character at " + x + ", " + y + " is " + character + " as expected.");
> + else
> + log ("FAIL: Character at " + x + ", " + y + " is " + actualCharacter + ". Expected " + character + ".");
Nit: extra space after log
Fixed in r167817.