Layout pixel snaps the position of segments for shape-inside. Modify the rounded rectangle tests to use the same pixel snapping as the polygon tests. In sub-pixel layout, the result may be slightly different.
Created attachment 171321 [details] initial patch
Created attachment 171511 [details] Factoring out subpixel test code to separate file
Modifying bug to cover factoring subpixel code out of polygon tests as well.
Created attachment 171518 [details] Adding polygon tests
Comment on attachment 171518 [details] Adding polygon tests Attachment 171518 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/14662059 New failing tests: fast/exclusions/shape-inside/shape-inside-rounded-rectangle-003.html
Comment on attachment 171518 [details] Adding polygon tests Can you investigate why one of the exclusion tests is failing please? Does subPixelLayoutEnabled() not help here? Is it another problem and we need to skip it?
Created attachment 171905 [details] Updating patch Forgot to pass along the subpixel argument when generating tests.
Created attachment 172099 [details] refactoring subpixel rounding into the .js file
Comment on attachment 172099 [details] refactoring subpixel rounding into the .js file LGTM. r=me.
Comment on attachment 172099 [details] refactoring subpixel rounding into the .js file Clearing flags on attachment: 172099 Committed r133384: <http://trac.webkit.org/changeset/133384>
All reviewed patches have been landed. Closing bug.