1933408 - [css-sizing-4] layout support for `stretch` and `-webkit-fill-available` keywords (preffed off by default)

Status: NEW

(Core :: Layout, enhancement)

Description

[Daniel Holbert [:dholbert]]

Spinning this bug off from bug 1789477 and bug 1872755.

This bug will contain the main layout patches to support these features (initially off-by-default), while those other bugs can remain open, pending pref-flips to enable the features.

Comment 1

[Daniel Holbert [:dholbert]]

Attachment: WIP: Bug 1933408 part 3: Implement layout support for 'stretch' as a block-axis size (with '-webkit-fill-available' as an alias).

This fixes the LinkedIn bug and also the gracobaby.ca bug, if you enable the
layout.css.webkit-fill-available.enabled about:config pref (added in an earlier
patch in this series).

Comment 2

[Daniel Holbert [:dholbert]]

Attachment: Bug 1933408 part 1: Rename several "bool stretch" vars to clarify that they're about the alignment keyword and not the sizing keyword. r?#layout

This is just a variable rename to help avoid ambiguity in upcoming patches.
This has no functional change.

Comment 3

[Daniel Holbert [:dholbert]]

Attachment: Bug 1933408 part 2: Introduce a "ComputeStretchContentBoxSize" helper to use in places where we're resolving something like the 'stretch' keyword. r?#layout

This patch doesn't change behavior[1]; it's just a refactor.

[1] Minor caveat: the "availBased" variables in this patch are now floored-to-0
whereas previously they were not. But we probably ended up effectively
clamping them (or their results) downstream from this point anyway; and in any
case, it seems like a step towards correctness to floor this size-measurement
to be nonnegative.

Comment 4

[Daniel Holbert [:dholbert]]

Attachment: Bug 1933408 part 4: Enable about:config prefs for WPTs, and update wpt expectations. r?#layout

Comment 5

[Daniel Holbert [:dholbert]]

Note: There are some frame state bits that I'm not worrying about for the purpose of this bug and have spun off to bug 1933554.