There are many changes in the RenderSVGShape area that may affect performance so we should add a performance test of hit testing to ensure we don't regress. Patch forthcoming.
Created attachment 151310 [details] Add a performance test for hit testing in SVG
Note: I have 5 groups of 12 shapes instead of 1 group and 5 <use>s because I hit the following regression in writing this test: https://bugs.webkit.org/show_bug.cgi?id=90820
Comment on attachment 151310 [details] Add a performance test for hit testing in SVG View in context: https://bugs.webkit.org/attachment.cgi?id=151310&action=review > PerformanceTests/SVG/SvgHitTesting.html:2 > +<!-- Benchmark of SVG Hit Testing --> Do we really need this comment? > PerformanceTests/SVG/SvgHitTesting.html:105 > + // var checkResults = {'ellipse': 0, 'circle': 0, 'rect': 0, 'path': 0, 'svg': 0, 'use': 0}; NIt: we should remove this. > PerformanceTests/SVG/SvgHitTesting.html:111 > + // checkResults[el.nodeName]++; Ditto. > PerformanceTests/SVG/SvgHitTesting.html:113 > + // console.log(checkResults); Ditto.
Created attachment 151729 [details] Update per reviewer comments
Comment on attachment 151729 [details] Update per reviewer comments Clearing flags on attachment: 151729 Committed r122351: <http://trac.webkit.org/changeset/122351>
All reviewed patches have been landed. Closing bug.