This function should measure complex text in the same manner as RenderBlock::LineBreaker::nextLineBreak().
Created attachment 104768 [details] Proposed changes. Note that I shamelessly stole the test from Mitz.
Comment on attachment 104768 [details] Proposed changes. Looks good but please run run-webkit-tests with the --pixel option and include the -expected.png file in your patch.
Created attachment 104772 [details] Added pixel results.
Comment on attachment 104772 [details] Added pixel results. Attachment 104772 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/9479039 New failing tests: fast/text/complex-preferred-logical-widths.html
(In reply to comment #4) > New failing tests: > fast/text/complex-preferred-logical-widths.html This is the test I just added. What does it mean?
Fixed in <http://trac.webkit.org/r93909>.
The test added by this patch is failing on Snow Leopard: http://build.webkit.org/results/SnowLeopard%20Intel%20Debug%20(Tests)/r93929%20(1875)/fast/text/complex-preferred-logical-widths-pretty-diff.html Does it just need a rebaseline?
Also failing on Windows: http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r94044%20(15853)/fast/text/complex-preferred-logical-widths-pretty-diff.html
(In reply to comment #7) > The test added by this patch is failing on Snow Leopard: > http://build.webkit.org/results/SnowLeopard%20Intel%20Debug%20(Tests)/r93929%20(1875)/fast/text/complex-preferred-logical-widths-pretty-diff.html > > Does it just need a rebaseline? The test had correct results and was passing as of r93909. Then incorrect results (which are incorrect for both Lion and Snow Leopard) were landed in <http://trac.webkit.org/r93924>.
(In reply to comment #9) > The test had correct results and was passing as of r93909. Then incorrect results (which are incorrect for both Lion and Snow Leopard) were landed in <http://trac.webkit.org/r93924>. :( +levin.
(In reply to comment #10) > (In reply to comment #9) > > The test had correct results and was passing as of r93909. Then incorrect results (which are incorrect for both Lion and Snow Leopard) were landed in <http://trac.webkit.org/r93924>. > > :( +levin. I don't know what happened there. Obviously a mistake -- yep I do them :). Feel free to revert it. I suspect it was an issue with the rebaselining tool that I didn't catch before submitting.
(In reply to comment #10) > (In reply to comment #9) > > The test had correct results and was passing as of r93909. Then incorrect results (which are incorrect for both Lion and Snow Leopard) were landed in <http://trac.webkit.org/r93924>. > > :( +levin. :) https://bugs.webkit.org/show_bug.cgi?id=67209