WebKit Bugzilla
Moving over test-cases from bug 87679 to focus on the writing-mode bug since that bug had two separate bugs (a flexbox bug and a writing-mode bug).
Created attachment 157545 [details] original testcase
Created attachment 157546 [details] reduction
As best I can tell, this is the correct behavior per the old spec: http://dev.w3.org/csswg/css3-writing-modes/#orthogonal-auto. But the definition of fill-available in the height direction has changed in the new spec to walk up the tree until it finds a containing block with a definite height: http://dev.w3.org/csswg/css3-sizing/#extrinsic-sizing.
(In reply to comment #3) > As best I can tell, this is the correct behavior per the old spec: http://dev.w3.org/csswg/css3-writing-modes/#orthogonal-auto. > > But the definition of fill-available in the height direction has changed in the new spec to walk up the tree until it finds a containing block with a definite height: http://dev.w3.org/csswg/css3-sizing/#extrinsic-sizing. There's no fill-available in your test case though?
(In reply to comment #4) > There's no fill-available in your test case though? The writing-mode spec link defines the auto-sized height for an orthogonal writing-mode in terms of the fill-available height.
Created attachment 162893 [details] Patch
Created attachment 163956 [details] Patch
Created attachment 163993 [details] Patch
Comment on attachment 163993 [details] Patch Attachment 163993 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/13839844 New failing tests: fast/writing-mode/percentage-height-orthogonal-writing-modes-quirks.html
Created attachment 164234 [details] Patch
Comment on attachment 164234 [details] Patch The table border models test case looks like a regression. "Header One" is not fitting on one line.
(In reply to comment #11) > (From update of attachment 164234 [details]) > The table border models test case looks like a regression. "Header One" is not fitting on one line. I'm pretty sure this is correct. If you load it at tip of tree and force the height of the vertical-rl div to 586px, then you get the same wrapping behavior. This patch causes the difference because we now subtract the body's margin when computing the availableHeight.
Created attachment 167780 [details] Update to ToT
Anyone mind taking a stab a reviewing this?
Created attachment 181335 [details] Patch
Comment on attachment 181335 [details] Patch r=me
Comment on attachment 181335 [details] Patch Clearing flags on attachment: 181335 Committed r138838: <http://trac.webkit.org/changeset/138838>
All reviewed patches have been landed. Closing bug.