|
tin
1.5.9
|
Verification and evidence artifacts make behavior, generated outputs, resource use, and replay coverage inspectable. They give a project a concrete trail from machine definition to generated reviews, host tests, target profiles, and resource manifests.
The evidence path is an engineering review workflow. It records what was checked, what was generated, and which bounded runtime surfaces were selected.
The current tooling and runtime surfaces use stable schema names for reviewable artifacts:
| Artifact | Schema | Use |
|---|---|---|
| Artifact manifest | tin.artifacts.v1 | target, runtime, transport, evidence, generated artifact, and package metadata index |
| Machine IR | verified-tsm.machine.v1 | state, event, transition, runtime, transport, and evidence metadata |
| Trace replay | verified-tsm.trace.v1 | finite event traces used by generated conformance tests |
| Traceability | verified-tsm.traceability.v1 | requirement-to-test and trace coverage mapping |
| Resource manifest | tsm.resource_manifest.v1 | static runtime resource accounting |
Runtime resource manifests record bounded storage and accounting information in a JSON form that can be checked into an evidence bundle.
The caller owns the stream and decides where the resulting bytes go.
Finite traces are the bridge between specifications, generated conformance tests, and runtime behavior checks. Trace generation records expected behavior as data, then emits C++ test cases that replay those traces against the public runtime surface.
The trace path is useful for:
The tooling treats traces as finite examples. It does not claim that a trace file proves every possible interleaving.
Run the quality build and tests before treating the documentation and generated artifacts as current:
The quality suite covers machine tooling smoke tests, generated Python import checks, generated pybind11 skeleton checks, trace generator checks, resource manifest checks, and runtime behavior tests.