2047061 - Should we skip empty line when computing inline containing block rect

Status: NEW

(Core :: Layout: Positioned, defect)

Description

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase #1: Inline containing block involves <br> and ib-split

Filed after reading Bug 2044288 Comment 6. This bug gathers more inline as abspos containing block rendering difference between Firefox and Google Chrome.

In ComputeInlineAbsoluteCBRect((), we simply grab the first and last inline continuations, and use them to compute the inline containing block rect. This behavior aligns with the current spec https://drafts.csswg.org/css-position-3/#absolute-cb

However, it seems Google Chrome will skip non-rendering <br> (or empty line?), and find the first (and/or last) non-empty inline frame or block frame to form the inline containing block rect. They might implement something that has been discussed in https://github.com/w3c/csswg-drafts/issues/8284, which is not yet reflected in the spec.

Webkit's rendering is more like us.

Comment 1

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Given the resolution in https://github.com/w3c/csswg-drafts/issues/8284 is not reflected in the spec, and Webkit's rendering is mostly aligned with us, the webcompat pressure is probably not high.

Daniel, what do you think? Any other thoughts?

Comment 2

[Daniel Holbert [:dholbert]]

Thanks for tracking down that spec discussion! I thought I recalled this coming up recently-ish.

The spec issue you linked indicates that there are still some details to be worked out. Given that we match Safari right now and we're not aware of many sites depending on Chrome's behavior right now, I'd recommend sitepatching the one known-affected site and waiting on the spec text to be finalized (or at least better hashed out) before starting work here.