[kgryte]

Resolves a part of #11352.

Description

What is the purpose of this pull request?

This pull request:

• migrates the tests for stats/base/dists/planck/skewness from relative tolerance testing to ULP difference testing, per [RFC]: Migrate math/base/special packages from relative tolerance testing to ULP difference testing (tracking issue) #11352.

Changes are confined to test/test.js and test/test.native.js. In each, the fixture loop's delta/tol computation and the y === expected[i] branch are replaced with a single isAlmostSameValue assertion, @stdlib/assert/is-almost-same-value is added to the module requires, and the now-unused @stdlib/math/base/special/abs and @stdlib/constants/float64/eps requires are removed. No other test cases, loop bounds, or fixtures are touched.

ULP bound: 1 in both test/test.js and test/test.native.js.

This is the measured minimum, not a guess. The JS and native implementations were each evaluated against all 1000 Python fixture cases, and the per-case ULP distance between the returned and expected values was computed directly via @stdlib/number/float64/base/ulp-difference. The resulting distribution was identical for both implementations: 891 cases at 0 ULP and 109 cases at exactly 1 ULP, so the maximum over the full fixture set is 1. The worst case is lambda = 7.28771354747294, where the returned 38.265183277421535 differs from the expected 38.26518327742153 by one representable step. Lowering the bound to 0 fails 109 of the 1000 cases, so 1 is the tightest integer bound that holds.

The C and JavaScript implementations return bit-identical results across all 1000 fixture cases, so a single bound is correct for both files.

The previous bound was 1.0 * EPS in both files (guarded by an exact-equality fast path). A 1 ULP bound is at least as tight everywhere, and strictly tighter for values just above a power of two, where 1.0 * EPS * |expected| spans nearly two representable steps.

Related Issues

Does this pull request have any related issues?

This pull request has the following related issues:

• [RFC]: Migrate math/base/special packages from relative tolerance testing to ULP difference testing (tracking issue) #11352

Questions

Any questions for reviewers of this pull request?

No.

Other

Any other information relevant to this pull request? This may include screenshots, references, and/or implementation notes.

Verification performed:

• make test TESTS_FILTER=".*/stats/base/dists/planck/skewness/.*" passes: 1005/1005 in test/test.js and 1005/1005 in test/test.native.js.
• The native addon was built locally (node-gyp rebuild) so that test/test.native.js actually exercised the C implementation rather than being skipped; it passes at the same 1 ULP bound. The build artifacts were removed before committing.
• The full suite was run twice at the final bound to confirm determinism (no FMA/architecture-dependent variation).
• The lower bound was confirmed empirically as well as analytically: temporarily setting the bound to 0 produces 109 failures, matching the measured ULP histogram exactly.
• Both files lint clean under etc/eslint/.eslintrc.tests.js, and pass editorconfig-checker under both etc/editorconfig-checker/.editorconfig_checker.json and .editorconfig_checker.markdown.json.
• Only the two test files are changed; no build artifacts are included.

Checklist

Please ensure the following tasks are completed before submitting this pull request.

• Read, understood, and followed the contributing guidelines.

AI Assistance

When authoring the changes proposed in this PR, did you use any kind of AI assistance?

• Yes
• No

If you answered "yes" above, how did you use AI assistance?

• Code generation (e.g., when writing an implementation or fixing a bug)
• Test/benchmark generation
• Documentation (including examples)
• Research and understanding

Disclosure

This PR was authored by Claude Code running as an automated task. The conversion mirrors the idiom established by previously merged stats/base/dists/* conversions (for example planck/stdev, rayleigh/variance, negative-binomial/stdev), and the ULP bound was determined empirically by measuring the per-case ULP distance across the full fixture set for both the JavaScript and C implementations, rather than by trial and error.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
stats/base/dists/planck/skewness	$\\color{green}162/162$ $\\color{green}+100.00\\%$	$\\color{green}8/8$ $\\color{green}+100.00\\%$	$\\color{green}2/2$ $\\color{green}+100.00\\%$	$\\color{green}162/162$ $\\color{green}+100.00\\%$

The above coverage report was generated for the changes in this PR.