docs(changelog): #7013 scalar-replacement root-slot hoist fragment

[proggeramlug]

PR #7013 (perf(gc): bind scalar-replacement root slots once per frame) was
squash-merged from a WIP commit before its changeset fragment was added. This
supplies the missing changelog.d/7013-*.md.

No code change — fragment only.

The fragment records what #7013 did and what backs it:

• Root cause. fix(gc): scalar-replaced object/array locals are not precise roots (#6968) #7007 emitted js_shadow_slot_bind at every store into a
  scalar-replaced field/element. A bind does four things, and for an
  entry-hoisted alloca three of them are loop-invariant — the address never
  changes, and every reader of a bound slot dereferences slot_ptrs[idx] in
  preference to the stack[idx] mirror the bind refreshes, so the mirror is
  dead storage. Only the incremental-mark shading barrier is genuinely
  per-store.

• Measured recovery, quiet Mac mini (load 1.9), pinned Node 26.5.0, 10M
  iterations, best-of-5, arms interleaved:

  arm	best
  pre-fix(gc): scalar-replaced object/array locals are not precise roots (#6968) #7007 (no rooting)	123 ms
  fix(gc): scalar-replaced object/array locals are not precise roots (#6968) #7007 as merged (bind per store)	164 ms
  perf(gc): bind scalar-replacement root slots once per frame, not once per store #7013 (bind per frame)	126 ms

  41 ms of rooting cost → ~3 ms, 93 % recovered. Binary size unchanged.

• The soundness obligation the hoist creates. Binding at entry makes the
  slot live from function entry, so the collector dereferences the alloca
  before any store reaches it. The array-element and fused-uppercase-receiver
  paths did not previously initialize their allocas; they now store undefined
  in entry_allocas like the object-literal path already did.

• Matrix. --arms all --pressure 8: PASS=229 UNVER=190 XFAIL=1 FAIL=20
  over 440 cells (22 rows, post-test(gc): de-duplicate the matrix corpus entry registered twice by #6977 #7011) — main's published 460-cell baseline
  minus exactly the row test(gc): de-duplicate the matrix corpus entry registered twice by #6977 #7011 de-duplicated (11 PASS + 9 UNVER). No cell
  changed state. Both FAIL files are the expected gc: a relocating minor with PRECISE roots breaks 14 of 20 representation-corpus files (5 crashes, 9 mismatches) — the conservative stack scan is load-bearing for correctness #6981 pair.

Summary by CodeRabbit

• Performance

  • Improved garbage-collection performance by reducing repeated work during memory updates.

• Bug Fixes

  • Strengthened memory-management correctness when values are initialized and tracked, preventing invalid data from being exposed.

• Documentation

  • Added changelog details covering performance results, correctness considerations, and validation outcomes.

• Tests

  • Expanded coverage for memory tracking, barrier behavior, and initialization requirements.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 7af1efe3-2ef8-4394-9d8e-134bb33be414

📥 Commits

Reviewing files that changed from the base of the PR and between df9c7d2 and 3d70bb4.

📒 Files selected for processing (1)

• changelog.d/7013-hoist-scalar-slot-root-binds.md

---

📝 Walkthrough

Walkthrough

The changelog documents hoisting scalar-replacement root slot bindings to frame entry, required initialization for previously unwritten allocas, GC validation results, relocation handling, and codegen test coverage.

Changes

Scalar slot root binding

Layer / File(s)	Summary
Hoisted binding and initialization contract changelog.d/7013-hoist-scalar-slot-root-binds.md	Documents entry-time binding, remaining per-store incremental-mark shading, and initialization requirements for hoisted slots.
Validation and test coverage changelog.d/7013-hoist-scalar-slot-root-binds.md, crates/perry-codegen/tests/scalar_replaced_slot_roots.rs	Records GC matrix results, relocation behavior, and tests covering binding placement and barrier behavior.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Possibly related PRs

• PerryTS/perry#7007: Adjusts scalar-replacement shadow-slot binding at store sites.
• PerryTS/perry#7013: Implements the scalar-replacement root-slot hoisting described by this changelog.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description covers the change, but it omits the required Summary/Changes headings, related issue, test plan, and checklist sections.	Add the template sections with a brief summary, concrete changes, related issue or n/a, test plan commands, optional screenshots/output, and the checklist items.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is concise and accurately describes the added changelog fragment for the scalar-replacement root-slot hoist.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch docs/7013-changelog-fragment

---

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.}