[Mikola Lysenko (mikolalysenko)]

LLM Description written by Claude Code:claude-opus-5-5

Sweep of crates/socket-patch-cli/src/commands/vendored_backend/{mod,repair}.rs, from the audit's coverage notes for these rows (living document: discussion #1830, Part 4 structure). Callers in vendor.rs (#1698), rollback.rs (#1864), remove.rs and commands/repair.rs (#1875) are held by open PRs, so the crate-visible API is unchanged and only file-local cleanups land here.

Changes

mod.rs

• Inlined revert_vendor_entry and deleted VendorRevertResult (its only caller was VendoredBackend::revert). The entry is now cloned once, so the separate flavor/gem pre-captures are gone.
• Added VendorRevertStep::unwired() in place of two hand-written match lists (the pypi advisory's list is unwired() || WouldRevert).
• Removed 7 bare issue references from comments.

repair.rs

• Inlined workspace_copy_issue (a forwarder to ArtifactHealth::is_workspace_copy_issue), rebuild_reason_label and format_ledger_missing.
• Candidate: dropped the unused Clone and the detached field, which copied entry.detached.
• One ensure_client replaces the two copies of the build-once API client code, and drops the expect.
• Tests: merged ui_format_tests into tests; renamed repair_returns_with_fifo_script → scan_skips_fifo_wiring_files (it tests the scan, not repair); removed the rebuild_reason_label passthrough assert along with the function. Removed 6 bare issue references from comments.

Size

git diff --stat: 2 files, +115 / −165. About +106 / −149 production and +9 / −16 test lines.

Merge with main

Merged main after #1954 to clear a merge-queue conflict: revert now passes the entry's wiring files (read from the cloned entry before the revert) to push_vendor_advisory, as main requires. Lib, in-process vendor/remove/rollback suites pass.

Behavior

None. Output bytes, JSON, exit codes and error codes are unchanged.

Verification

• cargo clippy --workspace --all-features -- -D warnings: clean.
• cargo test -p socket-patch-cli --lib: 940 passed.
• --test in_process_remove_repair_lifecycle, in_process_rollback_vendored, in_process_vendor, remove_rollback_api_overrides, remove, rollback: all passed.
• --test repair: 127 passed, 2 failed. The failures are the known root-sandbox chmod tests (repair_cleanup_failure_is_reported_in_json_and_silent_modes, repair_exits_zero_and_stays_quiet_when_lock_file_unremovable), which pass in CI.
• No cfg(windows)/cfg(unix) production code was changed.

/code-review high

It found no correctness bugs. I fixed 2 of its 8 low-severity findings: ensure_client no longer ends in an unwrap, and the scan tests' notes on the FIFO-safe reader and the one shared scan are restored. I left the other 6:

• remove.rs L1327's doc still names revert_vendor_entry, and commands/repair.rs has its own build-once client copy. Both files are held by Sweep remove.rs and repair.rs #1875; recorded in the register for the next sweep.
• The other => other arm in the preview label is unreachable, because a candidate's reason is always missing or corrupt. Making it an enum touches the event JSON reason field's source; it stays total.
• Inlining puts the step classification in the loop. That is acceptable: it had one caller, and the rules are documented on VendorRevertStep.
• The bare issue numbers were removed on purpose (the refactor procedure deletes unexplained issue numbers and keeps the why text).

Risk

Low. Mechanical inlining, covered by the in-process revert/repair suites.

🤖 Generated with Claude Code

https://claude.ai/code/session_011f9aePYVxtNghHHt1yrb1D

---

Note

Low Risk
Mechanical inlining and comment cleanup in vendored revert/repair paths; PR asserts unchanged behavior with broad CLI test coverage.

Overview
Refactors vendored_backend/mod.rs and repair.rs with no intended behavior change (same CLI output, JSON, and exit codes).

In mod.rs, revert_vendor_entry and VendorRevertResult are removed; their logic lives inline in VendoredBackend::revert, with one ledger entry clone up front instead of repeated lookups. VendorRevertStep::unwired() centralizes which revert outcomes count as “unwired” for PyPI reinstall advisories, gem stale-cache warnings, and go_undo_warnings. Bare GitHub issue numbers are stripped from comments while keeping explanatory text.

In repair.rs, small one-line helpers (workspace_copy_issue, rebuild_reason_label, format_ledger_missing) are inlined at call sites. Repair Candidate drops an unused Clone and redundant detached field (uses entry.detached). Shared lazy API client setup is factored into ensure_client, used by the redownload phase and fetch_record_by_uuid. Tests consolidate ui_format_tests into tests and rename the FIFO test to reflect that it exercises scan_vendor_references, not repair.

^{Reviewed by Cursor Bugbot for commit 0e72166. Configure here.}

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

BugBot review

---

Generated by Claude Code

[cursor]

Stale Bugbot comment from a previous run.

[Mikola Lysenko (mikolalysenko)]

Ready for review — burn-down agent.

• Head: 823cfbf6ee313f0ab593d3c29a77eb7619c65feb
• CI: all 82 check runs on this head succeeded or were skipped. None failed.
• Bugbot reviewed 823cfbf6 and found no issues.
• Mergeable with no conflicts. The PR doesn't touch CHANGELOG.md.
• For the reviewer: this is a refactor sweep of vendored_backend/mod.rs and repair.rs only. Not announced in Slack yet: the send tool wasn't available in this run, so a later run will retry.

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

Final review brief

What it does: A behavior-preserving cleanup of vendored_backend/mod.rs and repair.rs (+115/-165). It inlines single-caller helpers (revert_vendor_entry, rebuild_reason_label, format_ledger_missing, workspace_copy_issue), adds VendorRevertStep::unwired() in place of two hand-written variant lists, merges the two build-once API client blocks into ensure_client (dropping an expect), and strips bare issue numbers from comments.

Risk: low. It is a mechanical refactor, and every variant list, user-facing string and side-effect order matched the old code (checked below).

Look here:

• mod.rs:146-216:`` the inlined revert loop. Step branches, then entries.remove, then `save_state`, in the same order as before, with the same warning gating
• mod.rs:323-333:``` unwired() = `Reverted | Preserved | LedgerWriteFailed`. Its uses at :186 (`dry || unwired()`, the same as the old PyPI list) and :235 (the same as the old Go list)
• repair.rs:774-783:```` reads candidate.entry.detached in place of the removed field. `repair_workspace_copies` only changes `wiring`, so the value is the same
• repair.rs:886-911:````` ensure_client keeps build-once behavior, and the token advisory still prints at most once

Verified:

• Compared the old and new code for each inlined helper. The vendor_ledger_missing detail text is byte-identical after line continuation, and the offline guards are unchanged.
• cargo test -p socket-patch-cli --lib vendored_backend: 10 passed. cargo clippy -p socket-patch-cli --lib: clean.
• CHANGELOG.md is untouched and there is no debug code.
• CI: 31 checks passed, 20 skipped, 0 failed (ci-ok and clippy green). Cursor Bugbot passed. No open review threads.

Changes I made: none.

Nits (not blocking): remove.rs:1327 still names the deleted revert_vendor_entry (that file is held by #1875, and the PR description already defers this). The fall-through other => other reason label lost its one-line test.

Auto-merge (squash) is armed, so approving sends this straight to the merge queue.

---

Generated by Claude Code

[Wenxin Jiang (Wenxin-Jiang)]

[agent] Reviewed 823cfbf6ee313f0ab593d3c29a77eb7619c65feb. Checked vendored revert step transitions, advisory timing, ledger failures and lazy repair-client reuse. No major correctness issues found. Validation: source and regression-test review plus the available exact-head CI checks; no local test pass is claimed.

[Mikola Lysenko (mikolalysenko)]

BugBot review

---

Generated by Claude Code

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 0e72166. Configure here.}

[Mikola Lysenko (mikolalysenko)]

[final reviewer] Re-enqueued at 0e721669e: approved (approval stays valid after the post-approval pushes), ci-ok and clippy green, mergeable, no unresolved threads. Armed squash auto-merge, which put it back into the merge queue after the earlier dequeue.

---

Generated by Claude Code