Use tests/common's binary() and git_sha256 in 5 more CLI test files (#824)

[Mikola Lysenko (mikolalysenko)]

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

Refs #824 (children 2–3, slice 3)

Summary

Five more CLI test files drop their private binary() / git_sha256 and use tests/common's. After this change, finding the binary or hashing a blob is defined in one place for these files too.

Why

• Issue: #824. Register row C30 in register/20-audit-core.md. Living document: doc/08-tests-ci-docs.md (test-support bullet, {{C30}}).
• Leverage: B 0, U 0, D ≈7 private helper copies collapse onto tests/common (4 binary() and 3 git_sha256), R L (test-only). This was the highest-scoring candidate that no open PR blocks. Every higher-ranked production candidate (the manifest-load cluster An unparseable .socket/manifest.json is reported under five different --json error codes depending on the command #931…, Upstream restore sends every public-registry lookup at once; the registry_concurrency cap has no caller #1220, Delete the vendored and hosted-vlt helpers left without a production caller by the v5 consolidation #782/E41, Vendored requirements.txt refuses a six (==1.16.0) pin that the inventory and hosted mode accept, because exact pins are read by three grammars #1365, Lock inventory and the vendored PyPI router disagree on which lock governs when poetry.lock or pdm.lock has no packages #1114, The public-proxy per-package fallback runs 10 requests in flight, ignoring SOCKET_API_CONCURRENCY and the proxy cap of 4 #614/Chunk patch batch searches through one core helper with one set of batch limits #675, Validate patch UUIDs through one utils::uuid grammar instead of five #705) changes files that one of the 81 open PRs also changes.

What changed

• e2e_vex_redirect.rs, scan_rollout_e2e.rs and mode_migration_npm.rs declare #[path = "common/mod.rs"] mod common; and import binary (plus git_sha256 where they used it).
• get/global_packages_e2e.rs uses crate::common::binary. The get directory binary already declares mod common.
• vendor_ecosystem_fixtures/mod.rs uses crate::common::git_sha256, and its two includers (vendor_group_commit_e2e.rs, vendor_ledger_schema_e2e.rs) declare mod common.
• mode_migration_npm.rs takes cache_env/hermetic, and vendor_group_commit_e2e.rs takes envelope, through common. Before, they loaded the same files a second time through their own #[path] mods.
• cli/shared_helper_copies.rs removes the 5 migrated files from PENDING_PRIVATE_HELPERS.

Deleted

Production: +0/−0. Tests: +24/−50 across 8 files.

Behavior

None. No test changes what it asserts. common::git_sha256 is the canonical Git-blob hash, and common's oracle_selftests pin it to production's compute_git_sha256_from_bytes, which is the shape that both deleted kinds of copy had.

Test evidence

• cargo clippy --workspace --all-features -- -D warnings: clean.
• cargo test -p socket-patch-cli --all-features with these suites:
  • --test cli: 105 passed (includes the shared_helper_copies ratchet).
  • --test spawn_env_hygiene: 12 passed.
  • --test e2e_vex_redirect: 52 passed.
  • --test mode_migration_npm: 31 passed.
  • --test scan_rollout_e2e: 37 passed.
  • --test vendor_group_commit_e2e: 30 passed.
  • --test vendor_ledger_schema_e2e: 27 passed, 1 ignored.
  • --test get global_packages: 10 passed.
• Clippy over these test targets shows only warnings that already exist: the prebuilt_common lints and the common/jvm_env.rs duplicate-module warning, which cli and get emit on main too. CI's clippy job doesn't lint test targets.

Remaining (#824)

• Files changed by open PRs.
• The shared modules vex_e2e_common, vlt_e2e_common, vlt_hosted_common and vlt_vendor_common: not every includer declares mod common.
• Children 1, 4 and 5.

Risk

Low. Only test files change.

🤖 Generated with Claude Code

---

Note

Low Risk
Only CLI test sources and the helper-copy ratchet list change; runtime behavior and test assertions are unchanged.

Overview
Test-only refactor (#824): Five more CLI integration test targets stop defining local binary() / git_sha256() (and in some cases duplicate #[path] loads of cache_env, hermetic, or envelope) and wire through tests/common instead.

Affected binaries/modules include e2e_vex_redirect, scan_rollout_e2e, mode_migration_npm, get/global_packages_e2e, vendor_ecosystem_fixtures (plus vendor_group_commit_e2e / vendor_ledger_schema_e2e declaring mod common). Local SHA-256/blob hashing copies are removed in favor of common::git_sha256, which is pinned to production’s Git-blob hash.

The cli/shared_helper_copies ratchet list PENDING_PRIVATE_HELPERS drops those migrated paths so new private copies stay forbidden. No production or asserted test behavior changes—only shared helper usage and ~50 lines of duplicate test glue deleted.

^{Reviewed by Cursor Bugbot for commit d41d22f. Configure here.}

[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 d41d22f. Configure here.}

[Mikola Lysenko (mikolalysenko)]

[agent] ci-ok is red only because the CI workflow run on d41d22f (run 37978571474) was cancelled. Every non-green job reads cancelled, none reads failure, and no newer run superseded it. I re-ran that run once. This PR changes only test files, and the touched suites pass locally (listed in the description).

---

Generated by Claude Code

[Mikola Lysenko (mikolalysenko)]

[agent] yarn-berry 4.18.0 (ubuntu-latest) failed on 1cad3b2 in e2e_yarn_legacy_cachekey_refusal_build::yarn3_default_compression_cachekey8_refused_by_hosted_and_vendored. The cause is "corepack yarn@3.8.7 unavailable (SOCKET_PATCH_YARN_E2E_REQUIRED=1 forbids skipping)": the runner could not fetch yarn 3.8.7 through corepack. This PR doesn't touch that test or anything it includes, the other 33 tests in that binary passed, and the same job is green on main at 1438b1a. No code fix exists or is needed here. The run is still in progress, so GitHub refuses a re-run for now. I'll re-run the failed jobs once when the run completes.

---

Generated by Claude Code