Add ASCII flag to StringImpl
Created attachment 156066 [details] Patch
Comment on attachment 156066 [details] Patch Attachment 156066 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/13422412
Comment on attachment 156066 [details] Patch Attachment 156066 [details] did not pass win-ews (win): Output: http://queues.webkit.org/results/13411883
Comment on attachment 156066 [details] Patch Attachment 156066 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/13415614
Created attachment 156078 [details] Patch
Comment on attachment 156078 [details] Patch Attachment 156078 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/13404827
Comment on attachment 156078 [details] Patch Attachment 156078 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/13419485
Comment on attachment 156078 [details] Patch Attachment 156078 [details] did not pass win-ews (win): Output: http://queues.webkit.org/results/13411903
Comment on attachment 156078 [details] Patch marking r- for now as it doesn't compile
How did you measure performance impact (particularly on JSC)?
Created attachment 156953 [details] Patch
(In reply to comment #10) > How did you measure performance impact (particularly on JSC)? I ran the performance tests for the chromium build. There was no change. Once I have a stable mac build, I'll rerun the tests and post the results. Is there anything you had in mind that I should post in particular, or do the Dromaeo tests cover your concerns?
(In reply to comment #12) > (In reply to comment #10) > > How did you measure performance impact (particularly on JSC)? > > I ran the performance tests for the chromium build. There was no change. Once I have a stable mac build, I'll rerun the tests and post the results. Is there anything you had in mind that I should post in particular, or do the Dromaeo tests cover your concerns? Also run sunspider and V8.
Created attachment 157179 [details] sunspider and v8 performance comparison
Created attachment 158263 [details] Patch
Comment on attachment 158263 [details] Patch temporarily setting cq+ for perfalizer bot to pick up the patch
Comment on attachment 158263 [details] Patch clearing cq again
Comment on attachment 158263 [details] Patch next try for the perfalizer bot
Created attachment 158407 [details] Performance tests results for 158263
(In reply to comment #10) > How did you measure performance impact (particularly on JSC)? Okay all the performance numbers are in. The v8 and sunspider tests show no change, and the full performance suite shows more or less inconclusive results, which are to be expected as the change is small. Is there anything else you'd like to see?