Verification

Verification is layered: compile the public surface, run focused package tests, then execute the cross-target suite and smoke artifacts.

Fast loop

moon check --target native
moon test src/model --target native
moon test src/runner --target native
moon test src/report --target native
moon fmt --check

Cross-target matrix

moon test --target native
moon test --target js
moon check --target native
moon check --target js
moon info

Native covers process workers, monotonic timing, and replay. JS catches serialization and pure-policy portability regressions.

Artifact smoke tests

tmpdir=$(mktemp -d)
moon run src/cli --target native -- report \
  testdata/report/sample.jsonl "$tmpdir/report.html"
moon run src/cli --target native -- replay \
  testdata/replay/sample.jsonl --dry-run
test -s "$tmpdir/report.html"

The report fixture verifies JSONL parsing, scaling projection, escaping, and self-contained output. The replay fixture verifies artifact-version checks, command restoration, and timeout display without executing a process.

Evidence boundaries

• stats tests prove deterministic percentile bootstrap bounds and reject invalid configuration; they do not claim BCa or hierarchical resampling.
• runner tests prove setup/timing ordering, balanced blocks, worker timeout, and typed failure preservation; they do not prove a machine's timer quality.
• report tests prove Plot IR and escaping invariants; they do not guarantee browser-specific typography or CSS rendering.
• tune_gemm tests prove candidate validity and numerical checks; they do not prove that a candidate is fastest on every CPU or target.

Coverage and generated interfaces

moon coverage analyze
moon info

pkg.generated.mbti snapshots are the public-interface evidence. If a public signature changes, update the generated snapshot through MoonBit tooling and then update the matching package reference and localized index. Do not hand edit generated files.