absoluteClippedOverflowRect implemention does return an empty rect for offscreen nodes. so hasOffscreenRect (SpatialNavigation.cpp) can bail out earlier in such cases. see related discussion in webkit-dev in [1] [1] http://lists.macosforge.org/pipermail/webkit-dev/2010-March/012075.html
Created attachment 53405 [details] patch v1
Attachment 53405 [details] did not build on chromium: Build output: http://webkit-commit-queue.appspot.com/results/1591506
Attachment 53405 [details] did not build on gtk: Build output: http://webkit-commit-queue.appspot.com/results/1552523
Attachment 53405 [details] did not build on mac: Build output: http://webkit-commit-queue.appspot.com/results/1689241
Attachment 53405 [details] did not build on qt: Build output: http://webkit-commit-queue.appspot.com/results/1686255
Created attachment 53438 [details] (committed: r57663) patch v2 same as patch v1, but un-bitrotted and builds
Comment on attachment 53438 [details] (committed: r57663) patch v2 Clearing flags on attachment: 53438 Committed r57663: <http://trac.webkit.org/changeset/57663>
Thank you Simon Fraser!
Revision r57663 cherry-picked into qtwebkit-2.0 with commit bf61de5859ad29d41a76791862ac4c02b336d6da