CI: share test helpers in 17 more test files (#824)

[Mikola Lysenko (mikolalysenko)]

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

Refs #824 (slice 4 of the shared-test-helper migration; the issue stays open for the remaining files).

What it does

Test-only. Seventeen CLI test files still defined their own binary() and/or git_sha256 helpers. They now use tests/common's versions: top-level targets through #[path = "common/mod.rs"] mod common;, and the apply/, cli/ and rollback/ directory targets through use crate::common::... Imports that the deleted copies needed (sha2::{Digest, Sha256}, PathBuf) are dropped. scan_vendor_e2e.rs keeps use sha2::Digest because it still calls Sha512::digest.

cli/shared_helper_copies.rs (the ratchet that stops new copies) removes those 17 files from PENDING_PRIVATE_HELPERS, so any copy added back to them now fails the test. It also drops the stale diff_created_file_e2e.rs entry, because that file no longer exists.

Every copy removed here had one of the two shapes the ratchet's header says common pins: the Git-blob sha256("blob <len>\0" ++ content), or env!("CARGO_BIN_EXE_socket-patch"). So no test changes what it computes.

Still pending (unchanged)

• repair/covgap_commands_repair.rs, repair/covgap_commands_repair_vendor.rs, repair/repair_invariants.rs, scan/covgap_commands_fetch_stage.rs: open PR CI: hermetic spawns in scan/repair tests (#823) #1398 changes these files, so they stay out of this PR to avoid conflicts.
• vex_e2e_common, vlt_e2e_common, vlt_hosted_common, vlt_vendor_common: shared modules whose includers don't all declare mod common.

CI cost

No workflow change. A lean PR run is unchanged: same jobs and roughly the same job-minutes as before (about 28 Linux jobs, about 8 min wall time).

Validation

• cargo fmt --all -- --check: clean
• cargo test -p socket-patch-cli --all-features on all 17 touched targets (e2e_vex_vendor, the 11 in_process_* files, remove_rollback_api_overrides, scan_vendor_e2e, apply, cli, rollback): all pass with 0 warnings. That includes cli::shared_helper_copies, which checks the shrunken pending list. e2e_vex_vendor and scan_vendor_e2e have no #[ignore] tests.
• Not run locally: the whole cargo test --workspace, because the sandbox disk could not hold every test binary at once. Workspace clippy does not lint test targets. CI runs both.

/code-review high

No correctness findings. One cleanup fixed in 52739d9: `in_process_alternate_installers` and `in_process_gem_apply` compiled `common/cache_env.rs` twice (once on its own, once inside `mod common`). Both files now use `common::cache_env`, and both targets pass again. Findings left out on purpose:

• The ratchet matches only the names `binary()` and `git_sha256(`, so `git_blob_sha256` (2 files), inline `env!("CARGO_BIN_EXE_socket-patch")` (about 10 files) and `const BINARY` get past it. That needs a wider detector and more migrations, so it is a follow-up for Tracking: share CLI test helpers through one test-support module instead of 100+ per-file copies #824 and not this slice.
• `e2e_vex_vendor` and `remove_rollback_api_overrides` still build their own `SOCKET_*` scrub instead of using `common::hermetic`. That is Spawn every CLI test child through one hermetic Command builder; 10 test files inherit ambient SOCKET_* today #823's subject (CI: hermetic spawns in scan/repair tests (#823) #1398 covers scan/ and repair/), so it is a follow-up.
• `has_cargo` and `has_gem` hand-roll `common::has_command`. They behave slightly differently (`success()` vs `is_ok()`), so merging them is a behaviour change and a follow-up.
• `in_process_remove_repair_lifecycle` and `scan_vendor_e2e` include both `prebuilt_common` (which brings its own `jvm_env`) and `common`, so `jvm_env` selftests run twice there. Fixing that means changing the shared `prebuilt_common` module, which every other includer uses too, so it is out of scope.
• Each migrated top-level target now also compiles `common` and runs its few millisecond-scale selftests. This is the same trade the earlier slices (Use tests/common's binary() and git_sha256 in 40 more CLI tests (#824) #1258, Use tests/common's binary() and git_sha256 in 5 more CLI test files (#824) #1366) made.

🤖 Generated with Claude Code

https://claude.ai/code/session_017k2mexmgXxWiKdHrkRuRJC

---

Generated by Claude Code

[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 52739d9. Configure here.}

[Mikola Lysenko (mikolalysenko)]

Ready for review at 52739d9.

• CI: 36/36 non-skipped checks green (50 skipped).
• Bugbot: reviewed 52739d9, no findings.
• Mergeable: yes, no conflicts with main. Already approved by Tanmay Singla (@Tanmay182003) at this head.
• Slack announcement not sent: the Slack send tool was unavailable this run, so the next run will retry.

---

Generated by Claude Code