fast/repaint/selection-gap-fixed-child.html and fast/repaint/selection-gap-flipped-fixed-child.html have been failing on WK2 bots ever since they were added in http://trac.webkit.org/changeset/148258 http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Frepaint%2Fselection-gap-fixed-child.html%2Cfast%2Frepaint%2Fselection-gap-flipped-fixed-child.html
Hmm is anybody with a mac able to debug this? Looks like it's not invalidating the rect at all, when the selection is set. If you grab a pixel snapshot, is the box red or green?
Added test expectation in http://trac.webkit.org/changeset/148362.
Talked to Simon about this. In WebKit2, we become layer-backed on this test, so it's expected that there is no repaint logged. We could WontFix the test, or we could land the actual results. I'll go with the latter - the test is not testing much when layer-backed anyway, but at least we'll be catching crashes.
Landed the results in <http://trac.webkit.org/r171348>.