HayaoSuzuki/hoimin

Ran the README quick-start mutation test (hoimin run --root . --file src/calc.py --format human -- <test command>) on the pinned repo's basic Python fixture (tests/fixtures/projects/basic, copied to /tmp/hoimin-basic): baseline pytest passed, 3 mutants generated and all 3 killed, score 1.00, exit 0.

cat README* 2>/dev/null | head -200; ls -la; ls <host-path> 2>/dev/null
cd <host-path> && git log -1 --format=%H 2>/dev/null; ls -la; which node npm python3 uv pip3 2>&1; cat /etc/os-release | head -2
cd /tmp && (command -v uv || (curl -LsSf https://astral.sh/uv/install.sh | sh)) 2>&1 | tail -3; export PATH=$HOME/.local/bin:$PATH; uv --version; uv python install 3.14 2>&1 | tail -3
sed -n 1,80p <host-path>
export PATH=$HOME/.local/bin:$PATH; uv tool install --python 3.14 hoimin 2>&1 | tail -5; hoimin --help 2>&1 | head -40; which hoimin; ls -la ~/.local/bin | head
cd <host-path> && find tests tools -maxdepth 3 | head -50; grep -rn "calc.py\|allow-best-effort" docs/usage.md | head -20; grep -n "^## \|^### " docs/usage.md | head -60
cd <host-path> && find tests/fixtures/projects/basic -type f | head; sed -n 61,110p docs/usage.md; sed -n 40,60p docs/usage.md
rm -rf /tmp/hoimin-basic && mkdir -p /tmp/hoimin-basic && cp -r <host-path> /tmp/hoimin-basic/ && cd /tmp && uv venv -q -p 3.14 /tmp/hoimin-venv && uv pip install -q -p /tmp/hoimin-venv/bin/python pytest 2>&1 | tail -3; cd /tmp/hoimin-basic && /tmp/hoimin-venv/bin/python -m pytest -q 2>&1 | tail -3; cat /sys/fs/cgroup/cgroup.controllers 2>&1 | head -1
cd /tmp/hoimin-basic && timeout 600 <host-path> run --root . --file src/calc.py --format human -- /tmp/hoimin-venv/bin/python -m pytest -q 2>&1 | tail -40; echo "rc=${PIPESTATUS[0]}"
cd /tmp/hoimin-basic && timeout 900 <host-path> run --root . --file src/calc.py --format human --allow-best-effort-memory -- /tmp/hoimin-venv/bin/python -m pytest -q 2>&1 | tail -40; echo "rc=${PIPESTATUS[0]}"
cd <host-path> && mkdir -p .coldpass && ls -la .coldpass && cat > .coldpass/probe.json <<'EOF'
{
  "kind": "cli",
  "argv": ["<host-path>", "run", "--root", ".", "--file", "src/calc.py", "--format", "human", "--allow-best-effort-memory", "--", "/tmp/hoimin-venv/bin/python", "-m", "pytest", "-q"],
  "cwd": "/tmp/hoimin-basic",
  "expected_text": "score: 1.00"
}
EOF
cat > .coldpass/report.json <<'EOF'
{
  "outcome": "working",
  "operation": "Ran the README quick-start mutation test (hoimin run --root . --file src/calc.py --format human -- <test command>) on the pinned repo's basic Python fixture (tests/fixtures/projects/basic, copied to /tmp/hoimin-basic): baseline pytest passed, 3 mutants generated and all 3 killed, score 1.00, exit 0.",
  "operation_scope": "CLI mutation run only. The README's 'Plan and verify with an agent' workflow and the Rust build were not exercised. Full test/lint suites were not run per instructions. The run uses --allow-best-effort-memory because this guest lacks delegated cgroup v2 controllers, so memory is not hard-capped. hoimin is a CLI, so the MCP start/list_tools step does not apply.",
  "steps": [
    "curl -LsSf https://astral.sh/uv/install.sh | sh (uv 0.13.1; prerequisite for the README install)",
    "uv python install 3.14",
    "uv tool install --python 3.14 hoimin (installed hoimin 0.13.1)",
    "hoimin --help (verified)",
    "copied tests/fixtures/projects/basic to /tmp/hoimin-basic; uv venv -p 3.14 /tmp/hoimin-venv; uv pip install pytest",
    "baseline: python -m pytest -q (1 passed)",
    "hoimin run --root . --file src/calc.py --format human -- /tmp/hoimin-venv/bin/python -m pytest -q (first attempt refused, see deviations)",
    "hoimin run ... --allow-best-effort-memory -- /tmp/hoimin-venv/bin/python -m pytest -q (exit 0, score 1.00)"
  ],
  "deviations": [
    "README 'Quick start' command without --allow-best-effort-memory failed: 'portable resource limits require --allow-best-effort-memory: failed enabling delegated controllers with +memory +pids ... Permission denied'. Reran with --allow-best-effort-memory placed before -- as the README instructs for Linux without delegated cgroup v2.",
    "README's example target src/calc.py does not exist in the workspace; used the existing fixture tests/fixtures/projects/basic (src/calc.py + tests) copied to /tmp/hoimin-basic.",
    "Installed uv via its curl|sh installer and a Python 3.14 interpreter via uv, since the guest only has python3 3.12."
  ],
  "stuck_step": null,
  "main_blocker": null,
  "suggested_fix": null
}
EOF
python3 -m json.tool .coldpass/probe.json >/dev/null && python3 -m json.tool .coldpass/report.json >/dev/null && echo JSON_OK; <env-var>=<host-path> python3 -I <host-path> --validate 2>&1 | tail -20
grep -n "UNIT_SUITE\|useful product\|def " <host-path> | head -60; grep -n -B3 -A12 "UNIT_SUITE" <host-path> | sed -n 1,80p
sed -n 75,92p <host-path>; grep -n "agent_written(" <host-path>; cd <host-path> && cat tests/fixtures/projects/basic/pyproject.toml; ls tests/fixtures/projects/type-checking; grep -rn "unittest\|\"-m\"" docs/usage.md | head
cd /tmp/hoimin-basic && H=<host-path>; $H plan --help 2>&1 | sed -n 1,30p; echo ----; $H review --help 2>&1 | head -30; echo ----; $H progress --help 2>&1 | head -12
cd /tmp/hoimin-basic && timeout 300 <host-path> plan --root . --file src/calc.py --allow-best-effort-memory 2>&1 | head -c 1500; echo; echo "rc=${PIPESTATUS[0]}"
cd <host-path> && python3 - <<'EOF'
import json
p='.coldpass/report.json'
r=json.load(open(p))
r['stuck_step']='checker --validate rejects .coldpass/probe.json: "a unit suite does not establish a useful product operation" (the probe argv contains the token pytest, which is the README test command passed after -- to hoimin run)'
r['main_blocker']='The documented operation (hoimin run ... -- <project test command>) requires a test runner after --. The checker bans any argv token in {pytest, unittest, jest, vitest, mocha, tox, nox}, so no run-based probe passes validation. hoimin plan also requires a test argv after -- (verified: "at least one test argv element is required after `--`"), so it has the same conflict.'
r['suggested_fix']='Allow the test runner token when argv[0] is the product (for example, exempt hoimin run/plan/verify from the UNIT_SUITE rule), or accept an explicit check role for mutation-testing runs. Do not rename the runner to evade the rule.'
r['outcome']='working'
json.dump(r,open(p,'w'),indent=2)
EOF
python3 -m json.tool .coldpass/report.json >/dev/null && echo REPORT_OK; <env-var>=<host-path> python3 -I <host-path> --validate 2>&1 | tail -5; pgrep -fa "hoimin|pytest" | grep -v pgrep | head