See detail at http://crbug.com/126402. It seems that we should clear HarfBuzz's GSUB/GPOS feature settings every time.
Created attachment 140503 [details] Patch
Kent-san, could you take a look?
Comment on attachment 140503 [details] Patch ok
Comment on attachment 140503 [details] Patch Thank you!
Comment on attachment 140503 [details] Patch Rejecting attachment 140503 [details] from commit-queue. New failing tests: fast/text/international/bidi-AN-after-empty-run.html svg/W3C-I18N/text-anchor-dirLTR-anchorMiddle.svg svg/W3C-I18N/text-anchor-dirNone-anchorEnd.svg svg/W3C-I18N/text-anchor-dirRTL-anchorMiddle.svg fast/text/drawBidiText.html svg/W3C-I18N/text-anchor-dirRTL-anchorStart.svg accessibility/aria-disabled.html platform/chromium-linux/fast/text/international/draw-complex-text-from-to.html fast/frames/lots-of-objects.html css3/bdi-element.html fast/loader/text-document-wrapping.html fast/text/international/hindi-spacing.html fast/text/international/arabic-justify.html svg/W3C-I18N/g-dirLTR-ubNone.svg platform/chromium/virtual/gpu/fast/canvas/webgl/shader-precision-format.html fast/canvas/webgl/shader-precision-format.html svg/W3C-I18N/text-anchor-dirNone-anchorMiddle.svg fast/text/international/bidi-CS-after-AN.html fast/text/complex-text-opacity.html svg/W3C-I18N/g-dirLTR-ubOverride.svg css3/font-feature-settings-rendering.html http/tests/xmlhttprequest/xmlhttprequest-unsafe-redirect.html fast/text/international/complex-character-based-fallback.html svg/W3C-I18N/text-anchor-dirLTR-anchorEnd.svg fast/text/international/bidi-neutral-run.html svg/W3C-I18N/text-anchor-dirLTR-anchorStart.svg svg/W3C-I18N/text-anchor-dirNone-anchorStart.svg svg/W3C-I18N/text-anchor-dirRTL-anchorEnd.svg Full output: http://queues.webkit.org/results/12648232
Created attachment 140680 [details] Archive of layout-test-results from ec2-cq-02 The attached test failures were seen while running run-webkit-tests on the commit-queue. Bot: ec2-cq-02 Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'> Platform: Linux-2.6.35-28-virtual-x86_64-with-Ubuntu-10.10-maverick
Comment on attachment 140503 [details] Patch Attachment 140503 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/12651404 New failing tests: fast/frames/lots-of-objects.html fast/filesystem/workers/simple-temporary.html css3/font-feature-settings-rendering.html fast/text/international/arabic-justify.html css3/bdi-element.html
Created attachment 140768 [details] Archive of layout-test-results from ec2-cr-linux-04 The attached test failures were seen while running run-webkit-tests on the chromium-ews. Bot: ec2-cr-linux-04 Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'> Platform: Linux-2.6.35-28-virtual-x86_64-with-Ubuntu-10.10-maverick
I realized that the patch is wrong and the bug isn't easy to fix. I'll investigate later.