Plan: Regression coverage for conditional action dependency manifests (3.14.5)

[leynos]

Summary

Adds the execution plan for roadmap item 3.14.5 — Add regression coverage
for conditional action dependency manifests. This PR contains the plan only
(docs/execplans/3-14-5-regression-coverage-for-conditional-action-dependency-manifests.md);
no test or production code is included. The plan must be approved before
implementation begins.

The implementation it covers (3.14.2–3.14.4) already exists, so the plan scopes
test and documentation work only. It was produced with a planning agent
team, reviewed and revised by a community-of-experts agent team (structural,
contract/correctness, alternatives/DX, and reliability/ops lenses), and
cross-checked against the codebase. Tooling facts (googletest, pretty_assertions)
were researched via Firecrawl.

What the plan delivers

Regression coverage for all four roadmap sub-bullets:

1. Action-level when and foreach (integrated fixture; explicit action fan-out).
2. Complementary nextest-vs-legacy branches selecting exactly one action,
   driven through the real which resolver (existing complementary tests
   are stubbed).
3. Absent-command fallback without invoking shell(), asserted via
   StdlibState::is_impure() — the executable-discovery vs command-execution
   boundary.
4. deps lowering into IR implicit_deps and Ninja build statements, for
   conditionally-selected actions, including the {{ item }}-in-deps
   interaction.

Layers: in-crate rstest unit tests (selection / no-shell), IR integration
tests, insta Ninja snapshots validated with real ninja, and one combined
BDD scenario. Hexagonal framing pins the boundary without restructuring code.

Notable design decisions (see Decision Log)

• Injection uses the public StdlibConfig::with_path_override seam (not the
  private WhichConfig); in-crate placement is forced only by expand_foreach
  being pub(crate).
• is_impure() is honestly scoped: it flips for shell/grep/fetch.
• Property/bounded-model coverage of the same invariants is deferred to 4.2.x /
  4.3.2; ortho_config is out of scope.

Open questions for the approver

1. googletest mandate. The brief requires googletest, but the expert review
   unanimously recommended dropping it (keeping only pretty_assertions) given
   zero prior use, unverified #[gtest]+#[rstest] interop under the pinned
   rstest 0.18.0, and trivial assertions. The plan keeps it confined and
   spike-gated; the approver may instead choose pretty_assertions-only or
   googletest-throughout.
2. BDD depth — reduced to a single combined scenario; confirm or expand.
3. Stray src/stdlib/command/mod_backup.rs flagged for separate cleanup.

References

• Execplan: docs/execplans/3-14-5-regression-coverage-for-conditional-action-dependency-manifests.md
• Roadmap: docs/roadmap.md §3.14.5
• Lody session: https://lody.ai/leynos/sessions/29ac413e-35bb-41f6-8bdb-6b14d12e49a5

🤖 Generated with Claude Code

[sourcery-ai]

Sorry @leynos, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: e2f65de7-98e0-4f22-baef-36da49ee0fe0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch 3-14-5-regression-coverage-for-conditional-action-dependency-manifests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}