Add a microbenchmark for Hindi (complex-text) line-layout
Created attachment 169975 [details] Patch
I'm going to re-verify the slowdown with r131528 and then I'll file a regression bug if it's still slower than Safari 6.0.1.
I wrote this microbenchmark to make sure that bug 99393 wouldn't be a slowdown for complex text for Harfbuzz ports.
(In reply to comment #2) > I'm going to re-verify the slowdown with r131528 and then I'll file a regression bug if it's still slower than Safari 6.0.1. Filed as bug 100036.
Created attachment 170015 [details] the JS API changed out from under me
Ping reviewers?
Comment on attachment 170015 [details] the JS API changed out from under me View in context: https://bugs.webkit.org/attachment.cgi?id=170015&action=review > PerformanceTests/Layout/hindi-line-layout.html:36 > + function test() { > + style.display = "block"; > + style.width = "280px"; > + target.offsetLeft; > + style.width = "300px"; > + target.offsetLeft; > + style.width = "290px"; > + target.offsetLeft; > + style.display = "none"; > + } You can use anonymous function in the call to measureRunsPerSecond instead of defining a function since it’s not used elsewhere.
Comment on attachment 170015 [details] the JS API changed out from under me View in context: https://bugs.webkit.org/attachment.cgi?id=170015&action=review >> PerformanceTests/Layout/hindi-line-layout.html:36 >> + } > > You can use anonymous function in the call to measureRunsPerSecond instead of defining a function since it’s not used elsewhere. Thanks. This is just a copy of line-layout.html. If you'd like I can make a follow-up change to fix them both?
Comment on attachment 170015 [details] the JS API changed out from under me Clearing flags on attachment: 170015 Committed r132717: <http://trac.webkit.org/changeset/132717>
All reviewed patches have been landed. Closing bug.