It is helpful because we can say that MakeRope and ToString is side-effect free in this case!
Created attachment 307390 [details] Patch
Comment on attachment 307390 [details] Patch Attachment 307390 [details] did not pass jsc-ews (mac): Output: http://webkit-queues.webkit.org/results/3558116 New failing tests: stress/number-to-string-with-add.js.default stress/number-to-string-with-add.js.ftl-eager-no-cjit stress/number-to-string-with-add.js.ftl-no-cjit-validate-sampling-profiler stress/number-to-string-with-add.js.ftl-eager-no-cjit-b3o1 stress/number-to-string-with-add.js.dfg-eager-no-cjit-validate stress/number-to-string-with-add.js.ftl-no-cjit-no-put-stack-validate stress/number-to-string-with-add.js.ftl-no-cjit-no-inline-validate stress/number-to-string-with-add.js.ftl-eager stress/number-to-string-with-add.js.ftl-no-cjit-small-pool stress/number-to-string-with-add.js.no-cjit-collect-continuously stress/number-to-string-with-add.js.no-cjit-validate-phases stress/number-to-string-with-add.js.ftl-no-cjit-b3o1 stress/number-to-string-with-add.js.dfg-eager stress/number-to-string-with-add.js.dfg-maximal-flush-validate-no-cjit stress/number-to-string-with-add.js.no-ftl stress/number-to-string-with-add.js.no-llint
Created attachment 307393 [details] Patch
Comment on attachment 307393 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=307393&action=review r=me > Source/JavaScriptCore/ChangeLog:35 > + template_string_tag.es5 200.1027+-20.6871 ^ 25.7925+-11.4052 ^ definitely 7.7582x faster LOL bad benchmark :(
Committed r215472: <http://trac.webkit.org/changeset/215472>