As stated by http://www.w3.org/TR/css3-transitions/#transition-timing-function
Specifies a cubic-bezier curve. The four values specify points P1 and P2 of the curve as (x1, y1, x2, y2). Both x values must be in the range [0, 1] or the definition is invalid. The y values can exceed this range.
WebKit accepts "cubic-bezier(3, 0.1, 4, 1)" and does not reject the value as it should.
Created attachment 180501 [details]
Created attachment 180506 [details]
I removed the clamping in r96276 following discussion on www-style:
It seems like the specification still hasn't been updated, even though the last ED isn't that old. Dean, do you have any plans to update it, or have new concerns come up?
(In reply to comment #3)
> I removed the clamping in r96276 following discussion on www-style:
> It seems like the specification still hasn't been updated, even though the last ED isn't that old. Dean, do you have any plans to update it, or have new concerns come up?
Well if Dean confirms then I just need to update the tests to make sure we cover the behaviour correctly.
Yes, we should allow Y values to be > 1.
Created attachment 181189 [details]
Comment on attachment 181189 [details]
Clearing flags on attachment: 181189
Committed r138721: <http://trac.webkit.org/changeset/138721>
All reviewed patches have been landed. Closing bug.
Oops. We should clamp the X values though. It was my fault in the other bug:
The spec is correct AFAICT.
i.e. cubic-bezier(3, 0.1, 4, 1) should end up as cubic-bezier(1, 0.1, 1, 1)
Fixed by https://bugs.webkit.org/show_bug.cgi?id=106369. Closing the bug.