Consider merging https://chromium.googlesource.com/chromium/blink/+/2245e6b85adf3c96ef9b34dc6c92d736aaba2aac The resolution code in resolveGridPositionsFromStyle had a bug where we would check for 2 'auto' positions to determine if we needed to run the auto-placement algorithm. However the specification considers other positions as being treat as 'auto' depending on the context. Due to this glitch, we would try to explicitly resolve a 'span' position yielding to ASSERTs and the wrong result. While touching some grid parsing tests, added a real grid container so they now also test the resolution code path. Also removed some unneeded internal overrides. TEST=grid-item-bad-resolution-double-span.html
Created attachment 214132 [details] Patch
Committed r157389: <http://trac.webkit.org/changeset/157389>