Carry resolved anchor() insets on the abspos layout inputs

[kalenikaliaksandr]

Anchor resolution published its results through a pass-global side store
on LayoutState that every later inset style read consulted, running as
an implicit channel parallel to AbsposLayoutInputs, which already
crosses the run boundary into the out-of-flow sizing passes and is
persisted for replay. Make resolve_anchor_insets pure and thread its
result explicitly: the abspos engine stores it on the layout inputs and
hands it to the sizing chain, compute_inset consumes its own resolution
locally, and replay now reuses the resolutions the original pass saved
instead of reading unresolved anchor() values against an empty store.

This removes the last pass-global mutable state from layout:
LayoutState is now only the pass purpose.

[coderabbitai]

📝 Walkthrough

Walkthrough

The change adds computed wrapper handles for bare anchor() insets. It replaces the per-layout anchor inset store with ResolvedAnchorInsets. Absolute-positioned layout now returns resolved inset data and passes it through containing-block setup, sizing, root layout, replay, and relative inset computation. StyleValues applies resolved overrides, then wrapper payloads, then computed style values.

Sequence Diagram(s)

sequenceDiagram
  participant AbsposLayout as abspos_engine
  participant StyleValues
  participant ContainingBlock
  participant LayoutInputs as AbsposLayoutInputs
  AbsposLayout->>ContainingBlock: resolve anchor insets
  AbsposLayout->>StyleValues: read resolved inset values
  AbsposLayout->>LayoutInputs: save resolved_anchor_insets
  LayoutInputs-->>AbsposLayout: provide data during replay
  AbsposLayout->>StyleValues: apply inset overrides during sizing

Loading

Possibly related PRs

• LadybirdBrowser/ladybird#10912: Continues the style-payload migration used by ResolvedAnchorInsets.
• LadybirdBrowser/ladybird#10943: Refactors the anchor inset override machinery changed by this PR.
• LadybirdBrowser/ladybird#10961: Modifies the same anchor inset resolution and StyleValues paths.

Suggested reviewers: calme1709

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description check	✅ Passed	The description directly explains the removal of pass-global anchor-inset state and explicit threading of resolved insets, matching the changeset.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

🧹 Nitpick comments (1)

Libraries/LibWeb/Rust/src/layout/grid_formatting_context.rs (1)

3528-3530: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Tighten the last clause of the comment.

The first two claims are accurate. layout_pending_child() does resolve the insets later, at abspos_engine.rs Line 1728.

The third claim, "an anchor-bearing inset is never auto either way", holds only before resolution. Before resolution the side is InsetValue::BareAnchor, and is_auto() returns false. After resolution the side is InsetValue::Resolved, and is_auto() returns resolved.is_auto, which resolve_anchor_insets() sets to true when the anchor does not resolve. The recompute in base_containing_block_info() covers that case, so the code is correct; only the stated invariant is too strong.

📝 Proposed comment wording

-                    // Registration-time axis modes read raw style: anchor()
-                    // insets resolve later in layout_pending_child, and an
-                    // anchor-bearing inset is never auto either way.
+                    // Registration-time axis modes read raw style: anchor()
+                    // insets resolve later in layout_pending_child, where an
+                    // unresolvable anchor becomes auto and the axis modes are
+                    // recomputed from the resolutions.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@Libraries/LibWeb/Rust/src/layout/grid_formatting_context.rs` around lines
3528 - 3530, Update the comment near the axis-mode registration logic to limit
the “anchor-bearing inset is never auto” statement to the
pre-resolution/raw-style state. Mention that anchor insets may become auto after
resolution when the anchor cannot be resolved, while preserving the accurate
claims about later resolution in layout_pending_child().

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@Libraries/LibWeb/Rust/src/layout/grid_formatting_context.rs`:
- Around line 3528-3530: Update the comment near the axis-mode registration
logic to limit the “anchor-bearing inset is never auto” statement to the
pre-resolution/raw-style state. Mention that anchor insets may become auto after
resolution when the anchor cannot be resolved, while preserving the accurate
claims about later resolution in layout_pending_child().

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 6708c445-adc5-4f8c-a27f-d0ab33cfcefa

📥 Commits

Reviewing files that changed from the base of the PR and between 5b3e3fa and 3f96ded.

📒 Files selected for processing (7)

• Libraries/LibWeb/Rust/src/css/computed_value_types.rs
• Libraries/LibWeb/Rust/src/css/computed_values.rs
• Libraries/LibWeb/Rust/src/layout/abspos_engine.rs
• Libraries/LibWeb/Rust/src/layout/formatting_context.rs
• Libraries/LibWeb/Rust/src/layout/grid_formatting_context.rs
• Libraries/LibWeb/Rust/src/layout/layout_state.rs
• Libraries/LibWeb/Rust/src/layout/style_facts.rs