Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Test from leaf to top

Start focused

Run the owning crate and exact test target before the workspace:

cargo test -p hashsigs-types --locked -j1
cargo test -p hashsigs-reference --locked -j1
cargo test -p keccak-rhdl --locked -j1
cargo test -p sha256-rhdl --locked -j1
cargo test -p wots-rhdl --locked -j1

For WOTS changes, select the narrow integration test that owns the boundary: digits and blocks, task audit, context engine, transport, cluster, signer top, or verifier top.

Independent expectations

Expected hashes and signatures must come from an independent software oracle or pinned vector, not from a second call into the hardware logic under test. Control tests should use asymmetric tags and payloads so swaps and truncation do not accidentally pass.

Protocol cases

Exercise at least:

  • reset while work is live and post-reset quiescence;
  • valid held while ready is low;
  • bubbles and adjacent tokens;
  • full-width tag preservation;
  • saturation and a rejected extra offer;
  • malformed keep, last, identity, or generation;
  • stale, duplicate, missing, mistimed, and misrouted returns;
  • held result data under output backpressure; and
  • exact single credit release.

Expensive tests

Production full-lane source simulation and modular descriptor generation are ignored by default. Their evidence pages specify one-job concurrency, memory strategy, expected marker, and retained artifacts. Do not infer a pass because an ignored test was skipped.