Luna Flow
English

consistency tutorial

This page is for contributors: it explains how to run the cross-layer agreement tests and how to add one when you add or change an operation. Library users never import this package.

Quick start

Run the tests of this package from the repository (or from a workspace that contains it):

moon test -p Luna-Flow/luna-poly/consistency

All tests should pass; a failure means the two layers or two representations disagree.

Everyday tasks

Add a check for a new operation

When you add an operation to both layers, add a whitebox test to src/consistency/core_wbtest.mbt that computes it both ways and compares canonical outputs, for example:

test "new operation stays aligned" {
  let immutable = @immut.DensePolynomial::from_coefficients([1, 2, 3])
  let mutable = @mutable.DensePolynomial::from_coefficients([1, 2, 3])
  assert_true(
    immutable.new_operation().to_coefficients() ==
    mutable.new_operation().to_coefficients(),
  )
}

Compare through to_coefficients() or to_terms() (converting exponent vectors with to_array() if needed), because the two layers have different types.

Check a failure contract

For a checked API, assert that both layers return None on the same invalid input.

Going further

Laws of a single representation belong in that representation’s package (see src/immut/laws_wbtest.mbt, which uses moonbitlang/quickcheck); cross-representation and cross-layer equations belong here.

Common pitfalls

  • Comparing different types. @immut.DensePolynomial and @mutable.DensePolynomial are not comparable with ==; compare their canonical exports.
  • Order of to_terms(). Term storage is descending and sparse storage ascending; normalize through TermPolynomial::from_terms before comparing.

Next steps