Unblock the Rust Feature-Gate Smoke lane (embedding host seam in tests + tinymemory-core kernel-floor limit)

[YellowSnnowmann]

Summary

This PR unblocks the Rust Feature-Gate Smoke (gates off) lane, which was broken on main at four stacked layers. It makes four changes:

• Test seam — two migration-apply tests panic with "no EmbeddingHost installed" under --no-default-features. Fix: call the existing idempotent host_impls::install_for_tests() at the start of both.
• Kernel-floor ratchet — bump scripts/kernel-floor.limits from flows:305:282:2 to flows:307:284:2 to reconcile with the tinymemory-core + ureq client chain that became always-on via 6bf080266 (wired into openhuman_core on the always-on memory path without updating the ratchet). No new dependency is introduced here.
• CI dep-sim calibration — update .github/workflows/ci-lite.yml --expect-names 282 → 284 to match the new ratchet name count.
• Gated-test guard — remove openhuman/memory/people/address_book.rs from the feature-gated test-modules allowlist; the file was deleted from the tree, so the guard's EXPECTED list was stale.

Problem

migrate_openclaw_apply_imports_markdown_entries_into_target_workspace and migrate_hermes_apply_imports_markdown_entries run migrate_*(…, dry_run = false), whose apply path does memory work requiring tinymemory_core::embedding_host::set_embedding_host(...). Under the default feature set another test installs the process-global host first, so these passed by accident of ordering; under --no-default-features those tests are gated out and both panic.

Separately, the Feature-Gate Smoke lane also runs the kernel-floor ratchet, its dep-sim calibration, and a gated-test-modules guard. All three had drifted on main: tinymemory-core + ureq raised the flows profile from 305/282 to 307/284 without updating the ratchet or the --expect-names calibration, and address_book.rs was deleted while still listed in the guard's allowlist. These surfaced one after another as each earlier blocker was cleared.

Solution

1. Make the two apply tests self-sufficient by calling crate::openhuman::memory::host_impls::install_for_tests() (idempotent, Once-guarded).
2. Bump scripts/kernel-floor.limits to flows:307:284:2, with a history entry documenting the cause.
3. Update .github/workflows/ci-lite.yml --expect-names 282 → 284 to match.
4. Remove the stale openhuman/memory/people/address_book.rs entry from the gated-test allowlist.

Verified: GGML_NATIVE=OFF cargo test --no-default-features --lib -- apply_imports_markdown_entries → both tests pass; kernel-floor + dep-sim + gated-test guard all pass locally, and the Feature-Gate Smoke lane is green in CI. Linux resolves 307/284; macOS resolves 308/285 per the documented target skew.

Submission Checklist

• Tests added or updated — the two affected tests now install their own host seam; verified under --no-default-features.
• Diff coverage ≥ 80% — N/A: the Rust change is test setup (executes in the verified run); the other files are CI/ratchet config.
• Coverage matrix updated — N/A: no feature/behaviour change.
• All affected feature IDs from the matrix are listed — N/A.
• No new external network dependencies introduced — none; the ratchet/calibration reflect the graph already on main.
• Manual smoke checklist updated — N/A: test + CI-config change.
• Linked issue closed via Closes #NNN — N/A: infra fix; unblocks the Rust Feature-Gate Smoke lane (and Add lowercase openHuman launcher symlink to the .deb package #5513's CI).

Impact

• Unblocks the Rust Feature-Gate Smoke (gates off) lane for all PRs.
• The ops.rs change is test-only (no runtime/behaviour change).
• The ratchet bump and dep-sim calibration reflect the dependency graph already on main — no new dependency is added.
• The guard-list removal reflects a source file that no longer exists.

Related

• Unblocks CI for PRs that touch no Rust (e.g. Add lowercase openHuman launcher symlink to the .deb package #5513).

---

AI Authored PR Metadata (required for Codex/Linear PRs)

Linear Issue

• Key: N/A
• URL: N/A

Commit & Branch

• Branch: fix/migration-helpers-test-embedding-host
• Commit SHA: c0a2875

Validation Run

• GGML_NATIVE=OFF cargo test --no-default-features --lib -- apply_imports_markdown_entries → 2 passed
• cargo fmt --check — clean
• kernel-floor + dep-sim + gated-test guard — pass locally; Feature-Gate Smoke green in CI
• Full default-feature suite — N/A (heavy); the added call is an idempotent helper already used by default-build tests
• Tauri fmt/check — N/A (core crate + CI config only)

Validation Blocked

• command: full default-feature cargo test / clippy
• error: heavy local compile
• impact: none — verified under the exact failing config; CI is green on Feature-Gate Smoke

Behavior Changes

• Intended behavior change: none (test wiring + CI-config reconciliation)
• User-visible effect: none

Summary by CodeRabbit

• Tests

  • Improved memory migration test coverage for OpenClaw and Hermes, including standalone configurations.
  • Ensured test embedding support is available before migration scenarios run.

• Documentation

  • Added dependency measurement history for Linux and macOS environments.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: a0cbfcf4-358c-4160-9918-57f5eb4bf3f6

📥 Commits

Reviewing files that changed from the base of the PR and between c0a2875 and 28a2f48.

📒 Files selected for processing (2)

• scripts/kernel-floor.limits
• src/openhuman/config/migration_helpers/ops.rs

---

📝 Walkthrough

Walkthrough

The migration tests install the test embedding host before applying OpenClaw and Hermes migrations. The dependency-floor history records the tinymemory-core and ureq dependency chain and Linux/macOS measurements.

Changes

Migration test setup

Layer / File(s)	Summary
Install embedding host before migration src/openhuman/config/migration_helpers/ops.rs	The OpenClaw and Hermes apply migration tests install the test embedding host before migration. The setup supports standalone execution with --no-default-features.

Dependency floor history

Layer / File(s)	Summary
Record dependency measurements scripts/kernel-floor.limits	The history records the always-on tinymemory-core and ureq dependency chain, its effect on the flows profile, and Linux/macOS verification results.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related PRs

• tinyhumansai/openhuman#5502: Modifies the same migration tests to initialize test memory or embedding host seams.
• tinyhumansai/openhuman#5518: Installs the test memory or embedding host in the same migration tests and adds dependency-floor history.

Suggested labels: rust-core, bug

Poem

A rabbit sets the host in place,
Before migrations start their race.
OpenClaw hops, Hermes too,
Floor notes mark dependencies new.
Tests now burrow cleanly through.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
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.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the two main changes: embedding-host setup in tests and the tinymemory-core kernel-floor update for the Rust Feature-Gate Smoke lane.

---

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.}

[tinysweeper]

tinysweeper found nothing blocking. Approving.

             $0.0064 · 8,904 in / 1,029 out · 4,509 cached (51%) · z-ai/glm-5.2
critique:    $0.0018 · 2,015 in / 515 out   · 1,463 cached (73%) · z-ai/glm-5.2
security:    $0.0009 · 1,994 in / 116 out   · 1,460 cached (73%) · z-ai/glm-5.2
tests:       $0.0012 · 1,909 in / 295 out   · 1,522 cached (80%) · z-ai/glm-5.2
description: $0.0024 · 2,986 in / 103 out   · 64 cached (2%)     · z-ai/glm-5.2

[YellowSnnowmann]

Added a second commit (bf83c109): the Feature-Gate Smoke lane runs two checks, and the first commit only fixed one.

• Commit 1 (a6ec0bc9) — the --no-default-features test failures (no EmbeddingHost installed), fixed by installing the host seam in the two apply tests. Verified passing locally.
• Commit 2 (bf83c109) — the kernel-floor ratchet in the same lane also regressed on main: the flows profile resolves 307 packages / 284 names against a 305/282 limit. The growth is tinymemory-core + its ureq HTTP client chain (+2/+2), wired into openhuman_core on the always-on memory path by 6bf080266 without updating the ratchet. Raised the limit to 307/284, with the justification recorded in the file's history block (its own policy for a raise). No dependency is added — this only reconciles the ratchet to the graph already on main.

Both are required for the lane to go green. The macOS/Linux skew is expected (macOS resolves 308/285; the ratchet is calibrated on Linux at 307/284, which is what CI reported).

[tinysweeper]

What this change touches

2 files, +20 -0 across 2 components. The code graph knows nothing about these files yet — normal for newly added files, and a cold index otherwise.

flowchart LR
  n0["scripts<br/>1 file +12 -0"]:::changed
  n1["src/openhuman/config/migration_helpers<br/>1 file +8 -0"]:::changed
  classDef changed fill:#0d4429,stroke:#238636,color:#e6edf3
  classDef impacted fill:#161b22,stroke:#6e7681,color:#c9d1d9
  classDef flagged fill:#5a1e02,stroke:#d93f0b,color:#ffffff
  classDef blocking fill:#67060c,stroke:#f85149,color:#ffffff

Loading

_{Green: changed. Grey: untouched, reached through an import or a call. Orange: has findings. Red: has a finding that blocks the merge.}

Component		Files	Lines	Findings
scripts	changed	1	+12 -0	—
src/openhuman/config/migration_helpers	changed	1	+8 -0	—

Changed files

scripts

• scripts/kernel-floor.limits

src/openhuman/config/migration_helpers

• src/openhuman/config/migration_helpers/ops.rs

[tinysweeper]

Requesting changes: 1 lane(s) blocking, worst finding is high.

Fix or reply to the findings below and push. The next review clears this automatically once they are gone — you should not need to dismiss anything by hand.

             $0.0209 · 11,388 in / 7,409 out · 8,621 cached (76%) · z-ai/glm-5.2
critique:    $0.0036 · 2,187 in  / 1,285 out · 1,722 cached (79%) · z-ai/glm-5.2
security:    $0.0036 · 2,166 in  / 1,214 out · 1,536 cached (71%) · z-ai/glm-5.2
tests:       $0.0037 · 2,960 in  / 1,222 out · 2,269 cached (77%) · z-ai/glm-5.2
description: $0.0099 · 4,075 in  / 3,688 out · 3,094 cached (76%) · z-ai/glm-5.2

[tinysweeper]

The previously-blocking findings are resolved. Clearing the changes request.

             $0.0317 · 20,916 in / 10,702 out · 15,326 cached (73%) · z-ai/glm-5.2
critique:    $0.0110 · 6,880 in  / 3,613 out  · 4,530 cached (66%)  · z-ai/glm-5.2
security:    $0.0090 · 6,817 in  / 3,088 out  · 5,657 cached (83%)  · z-ai/glm-5.2
tests:       $0.0088 · 2,980 in  / 3,346 out  · 2,165 cached (73%)  · z-ai/glm-5.2
description: $0.0029 · 4,239 in  / 655 out    · 2,974 cached (70%)  · z-ai/glm-5.2