[css-text-decor-4] Computed value of text-decoration-thickness · Issue #10044 · w3c/csswg-drafts (original) (raw)

According to the current draft, the computed value of text-decoration-thickness is "specified keyword or absolute length", which implies that percentages must be resolved to lengths in the computed value.

This seems to be at odds with the note just below, which says that

A percentage will inherit as a relative value, and will therefore scale with changes in the font as it inherits.

If percentages were retained in the computed value, this would occur naturally, but with the computed value being an absolute length, inheritance seems to be a special case.

So was the definition of the computed value just overlooked when adding percentage support? AFAICS the current WPT test at https://wpt.live/css/css-text-decor/text-decoration-thickness-computed.html expects to see percentages retained.

On the other hand, there's an animation test at https://wpt.live/css/css-text-decor/animations/text-decoration-thickness-interpolation.html that expects to see percentages resolved to absolute px lengths. If the computed-value test is correct and percentages should be retained, then shouldn't they also be interpolated as such?