The tcheck2 test worked for the old JPZC because it used specific APIs to record checkerboarding that worked with JPZC. With the C++ APZ we have async subframe scrolling so the notion of checkerboarding is more complicated. We have plans to get detailed checkerboarding metrics in bug 1226826, and once we have that we should rewrite the tcheck2 test to use it. Fundamentally I think the tcheck2 test should be written the same way it was before (i.e. load a page in Fennec, simulate a bunch of user actions, measure checkerboarding), but perhaps expand it to cover different kinds of sites.
You need to log in before you can comment on or make changes to this bug.