# Changelog

## Score

- CAI 65 → 64 (-0.6)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.3) — scores are not directly comparable.

## Lenses

- Code Health 95 → 95 (+0.0)
- Architecture 98 → 98 (+0.0)
- Maturity 53 → 53 (+0.0)
- Readiness 67 → 69 (+1.7)
- Security 74 → 68 (-6.5)

## Resolved (5)

- Coverage not measured — no coverage collector is wired up
- Duplicate intent across multiple namespaces/types. The `equal` matcher/constraint is exposed identically in `MsTest`, `CustomMatchers`, `Xunit`, and `TopLevelOperators`. While these may target different test frameworks, the naming is identical without clear differentiation in the signature (e.g., no framework-specific suffix), leading to confusion about which one to use. Furthermore, `FsUnitTyped.TopLevelOperators` provides `shouldEqual` which is a direct functional duplicate of the assertion logic provided by `equal` in the other modules.
- Parameter order inconsistency. In `MsTest`, `CustomMatchers`, and `Xunit`, the signature is `equalWithin(tolerance, expected)`. In `TopLevelOperators`, the signature is `equalWithin(tolerance, expected)`. While they look the same, the `TopLevelOperators` version returns a `EqualConstraint` while others return `CustomMatcher`. This suggests a fundamental difference in implementation strategy (constraint-based vs matcher-based) that is not reflected in the naming, potentially confusing users who expect interchangeable usage.
- Redundant and inconsistent naming for length assertions. `haveLength` is exposed in `MsTest`, `CustomMatchers`, `Xunit`, and `TopLevelOperators`. Additionally, `FsUnitTyped` exposes `shouldHaveLength`. The parameter name changes from `n` in `MsTest` to `expected` in others, which is a minor inconsistency but reflects a lack of standardization. More importantly, `haveCount` exists alongside `haveLength` in the same modules, which is often redundant for collections where length and count are synonymous, or confusingly distinct for arrays vs lists.
- Signature inconsistency for sequence equality. `CustomMatchers.equalSeq` takes two arguments `(func, expected)`, whereas `MsTest`, `Xunit`, and `TopLevelOperators` versions take one argument `(expected)`. This breaks the pattern of consistent operator signatures across the library's different entry points.

## New (1)

- Edited copy of a member (145 corresponding lines) (src/FsUnit.MsTestUnit/FsUnit.fs)

## Changes since last survey

- 9 commits — 9 feature/other, 0 fixes

## By area

- .config/dotnet-tools.json — 2 commits
- src/FsUnit.MsTestUnit — 2 commits
- src/FsUnit.Xunit — 2 commits
- tests/FsUnit.MsTest.Test — 2 commits
- tests/FsUnit.Xunit.Test — 1 commit

## Notable commits

- change: Bump MSTest.TestAdapter and MSTest.TestFramework (#320)
- change: Bump MSTest.TestAdapter and MSTest.TestFramework (#329)
- change: Bump MSTest.TestFramework from 4.5.1 to 4.5.2 (#322)
- change: Bump MSTest.TestFramework from 4.5.2 to 4.5.3 (#331)
- change: Bump dotnet-outdated-tool from 4.8.1 to 4.9.0 (#327)
- change: Bump fantomas from 8.1.0 to 8.1.1 (#323)
- change: Bump xunit.v3.assert from 4.0.2 to 4.0.3 (#324)
- change: Bump xunit.v3.extensibility.core from 4.0.2 to 4.0.3 (#325)
- change: Bump xunit.v3.mtp-v2 from 4.0.2 to 4.0.3 (#326)
