Wrong expectations for percentage interpolation in text-decoration-thickness-interpolation.html

[jfkthame]

This test, added in 708e132, includes examples of interpolating text-decoration-thickness involving percentage values.

The test appears to expect all the percentages to be resolved to absolute pixel dimensions, but AFAICT this is incorrect. When interpolating between percentage values, the result should also be computed as a percentage; and when interpolating between a mix of length and percentage, the spec calls for:

...converting both values into a calc() expression representing the sum of the dimension type and a percentage (each possibly zero) and interpolating each component individually...

So it seems wrong to expect all the interpolations involving percentages to resolve to absolute dimensions.

This is currently resulting in a bunch of "failures" in Firefox of the form

CSS Transitions: property from [100%] to [0%] at (0) should be [16px]

and so on, where Firefox (correctly, I think) returns 100%, not 16px.

https://drafts.csswg.org/css-text-decor-4/#text-decoration-thickness-property
https://drafts.csswg.org/css-values-4/#combine-mixed

cc @shivamidow