Bug 237628 – [css-grid] Subgrid items should always be stretched

• Status: RESOLVED
• Resolution: FIXED
• Component: Layout and Rendering

[Matt Woodrow]

9.j: The subgrid is always stretched in its subgridded dimension(s): the align-self/justify-self properties on it are ignored, as are any specified width/height constraints.

We need to make sure that we always stretch subgrid items, even if it means overriding a specified width/height.

[Radar WebKit Bug Importer]

<rdar://problem/89996514>

[Matt Woodrow]

Created attachment 454186 [details]
Patch

[Matt Woodrow]

Created attachment 454188 [details]
Patch

[EWS Watchlist]

This patch modifies the imported WPT tests. Please ensure that any changes on the tests (not coming from a WPT import) are exported to WPT. Please see https://trac.webkit.org/wiki/WPTExportProcess

[Tim Nguyen (:ntim)]

Comment on attachment 454188 [details]
Patch

Upstream WPT doesn't seem happy about some trailing whitespace in your test, can you remove them?

See: https://community-tc.services.mozilla.com/tasks/fKANgyH9StmuqxCzp54LiA/runs/0/logs/public/logs/live.log#L108-110

[zalan]

Comment on attachment 454188 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=454188&action=review

> Source/WebCore/rendering/RenderBox.cpp:2885
> -        return style.resolvedJustifySelf(&containingBlock->style(), containingBlock->selfAlignmentNormalBehavior(this)).position() == ItemPosition::Stretch;
> -    return style.resolvedAlignSelf(&containingBlock->style(), containingBlock->selfAlignmentNormalBehavior(this)).position() == ItemPosition::Stretch;
> +        return resolvedJustifySelf(&containingBlock->style(), containingBlock->selfAlignmentNormalBehavior(this)).position() == ItemPosition::Stretch;
> +    return resolvedAlignSelf(&containingBlock->style(), containingBlock->selfAlignmentNormalBehavior(this)).position() == ItemPosition::Stretch;

I think it's ok to just check/downcast to RenderGrid and run the grid specific code in here. While I personally dislike the way we structure all these different layout systems, I think that's what the current pattern here is (same for width)

> Source/WebCore/rendering/RenderGrid.cpp:1208
> +    GridTrackSizingDirection childFlowDirection = GridLayoutFunctions::flowAwareDirectionForChild(*this, child, direction);

auto?

> Source/WebCore/rendering/RenderGrid.cpp:1292
> +{

I'd just early return on a non-grid child.

> Source/WebCore/rendering/RenderGrid.cpp:1295
> +        GridTrackSizingDirection childBlockDirection = GridLayoutFunctions::flowAwareDirectionForChild(*this, child, ForRows);
> +        LayoutUnit stretchedLogicalHeight = availableAlignmentSpaceForChildBeforeStretching(GridLayoutFunctions::overridingContainingBlockContentSizeForChild(child, childBlockDirection).value(), child, ForRows);

autos

[Matt Woodrow]

Created attachment 455023 [details]
Patch for landing

[Matt Woodrow]

Created attachment 455039 [details]
Patch for landing

[EWS]

Committed r291464 (248580@main): <https://commits.webkit.org/248580@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 455039 [details].