[kgryte]

Resolves a part of #11352.

Description

What is the purpose of this pull request?

This pull request:

• migrates the relative tolerance test cases in stats/base/dists/triangular/pdf to ULP difference testing.

The only file in this package still using relative tolerance was test/test.factory.js; the sibling test/test.pdf.js and test/test.native.js had already been migrated, so this change brings test.factory.js in line with them.

Each of the three fixture loops (small, medium, and large range) previously used:

if ( y === expected[i] ) {
    t.strictEqual( y, expected[i], ... );
} else {
    delta = abs( y - expected[ i ] );
    tol = 1.0 * EPS * abs( expected[ i ] );
    t.ok( delta <= tol, 'within tolerance. ...' );
}

and now uses:

t.strictEqual( isAlmostSameValue( y, expected[ i ], 0 ), true, 'returns expected value' );

The now-unused abs and EPS requires, along with the delta and tol declarations, were removed, and @stdlib/assert/is-almost-same-value was added.

ULP constant: 0 (all three fixture loops).

This is the measured minimum. The maximum observed ULP difference between the returned and expected values was 0 across all 3000 fixture values (1000 each for small_range, medium_range, and large_range), for both the factory and the direct pdf code paths, so no non-zero ULP allowance is required. Note that isAlmostSameValue with a maxULP of 0 falls back to a SameValue comparison; none of the fixtures contain a zero or signed-zero expected value, so this is safe. The implementation performs only plain IEEE 754 multiplication, subtraction, and division, which are correctly rounded and not subject to FMA contraction in JavaScript, so the result is bit-reproducible.

The package's test suite was run twice at the final ULP value to confirm determinism (3027 passing assertions in test.factory.js each time, with no failures).

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?

The already-migrated sibling test/test.pdf.js asserts the same three fixture sets with a ULP value of 1, where 0 would also pass. This pull request uses the measured minimum of 0 for test.factory.js, per the guidance in #11352 to use the minimum required ULP value. Happy to use 1 instead if you would prefer consistency with test.pdf.js over the tightest bound.

Other

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

• Only test files were changed; no implementation, documentation, or fixture changes.
• make test TESTS_FILTER=".*/stats/base/dists/triangular/pdf/.*" passes. test/test.native.js reports 0 tests in this environment because the native add-on was not built; this is pre-existing and unrelated to this change.
• make eslint-tests TESTS_FILTER=".*/stats/base/dists/triangular/pdf/.*" passes cleanly.
• The lint-editorconfig-files check could not be run in this environment, as it downloads the editorconfig-checker binary from GitHub and network access to that repository was unavailable. The changed lines were verified manually to use tab indentation, LF line endings, no trailing whitespace, and a final newline.

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

If you answered "yes" to using AI assistance, please provide a short disclosure indicating how you used AI assistance.

This pull request was authored by Claude Code, run as an automated task by @kgryte. The migration mirrors the already-migrated test/test.pdf.js in the same package, and the ULP bound was measured empirically against all 3000 fixture values rather than guessed.

---

@stdlib-js/reviewers

---

Generated by Claude Code