fix #23: [14] Guard against uncollected nested test functions

[CuriosityQuantified]

Closes #23

Summary

Issue #23 ([14] Nested test function is unreachable dead code) reported a test_* function defined inside test_create_idea_live_groq_call in tests/unit/test_llm_service.py. Pytest only collects module-level functions and Test*-class methods, so the nested copy was never run — dead code that masqueraded as fallback-parsing coverage.

The literal deletion already landed via PR #46 (commit d275ebf, which closed #22) as a side-effect of the same file's mock fix, so there is nothing left to delete. What was still missing is a guard against recurrence. This PR adds that guard, following the repo's per-issue regression convention (tests/unit/test_<slug>_regression.py + a named CI step).

Changes

• NEW tests/unit/test_nested_test_discovery_regression.py — AST-based, fully offline guard (it parses files with ast, never imports/execs them). It scans every tests/unit/test_*.py and fails if any test_* function is nested inside another function. A ClassDef body resets the enclosing-function context, so Test*-class methods stay allowed. Two targeted assertions lock in the exact [14] Nested test function is unreachable dead code #23 fix:
  • test_create_idea_final_turn_fallback_parsing is defined exactly once at module level;
  • it is not nested inside test_create_idea_live_groq_call.
• .github/workflows/ci.yml — one named step Run nested test discovery regression in the unit-tests job, mirroring the sibling per-issue regression steps. Deselect list and other jobs untouched.

Verification

• New regression: 3 passed. Verified it bites — injecting a nested def test_x inside def test_wrapper makes it fail with a precise message, then passes again once removed.
• Full unit suite (CI deselects): 231 passed / 1 skipped / 6 deselected (baseline was 228; +3 new).
• Next.js production build (npm run build): green.
• Playwright e2e job: unaffected — this diff touches no frontend/e2e code.

Pipeline

Four sequential fresh subagents ran in the shared checkout: implementation+CI, code review (APPROVE), simplification (removed one trivial single-use helper), and security review (LOW / clean — no material findings). No graphify-out/ or code-graph CI job exists in this repo, so graph refresh is not applicable.

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
save-the-world	Ready	Preview	Aug 10, 2026 1:04am
sim-local	Ready	Preview	Aug 10, 2026 1:04am