These reftests fail with what looks like antialiasing differences on Mac.
The failures appear to be WebKit2 only.
Updated expectations in <http://trac.webkit.org/r152470>.
hmm this caused those tests to fail on Lion Debug WK1 - not sure why that configuration is different from the rest of the WK1 tests or why the Release tests aren't failing.
Does the Lion failure look like color profiles, or antialiasing? Seems appropriate to skip on wk1 again with an appropriate comment and bug link.
Especially if it's a color profile difference, this could depend on bot hardware, or perhaps even on whether it's connected to a display at the moment.