[kgryte]

Resolves a part of #11352.

Description

What is the purpose of this pull request?

This pull request:

• migrates the tests for blas/base/cscal from relative tolerance testing (delta <= rtol * EPS * abs( expected )) to ULP difference testing using @stdlib/number/float32/base/assert/is-almost-same-value.
• replaces the EPS/abs imports with isAlmostSameValuef and introduces the private isApprox test helper which accepts a maximum allowed ULP difference, matching the idiom used in already migrated sibling packages (e.g., blas/base/wasm/cscal, blas/base/csrot, and blas/base/caxpy).
• tightens each assertion to the minimum ULP value which still passes over the full set of test cases.

Files changed (test files only):

• test/test.cscal.js
• test/test.cscal.native.js
• test/test.ndarray.js
• test/test.ndarray.native.js

test/test.js contains no tolerance-based comparisons and is left unchanged.

Measured minimum ULP bounds (per test case, in file order):

File	ULP values
test/test.cscal.js	1, 1, 1, 2
test/test.cscal.native.js	1, 1, 1, 2
test/test.ndarray.js	1, 1, 1, 5, 5
test/test.ndarray.native.js	1, 1, 1, 5, 5

These values coincide exactly with those used in the already migrated blas/base/wasm/cscal, which exercises the same inputs and expected values.

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 native test suites could not be executed locally, as the native add-on was not built in the development environment (the suites were skipped via tryRequire; building requires gfortran, which was unavailable). Following the convention in previously migrated packages, the native test files use the same per-test-case ULP values as their JavaScript counterparts. Please flag if the C implementation is expected to diverge for any of these cases.

Other

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

How the bounds were determined:

• The maximum observed ULP difference was measured for every test case over the full set of expected values using @stdlib/number/float32/base/ulp-difference.
• Each bound was then confirmed to be minimal: at N, the suite passes; at N-1, the corresponding test case fails.
• The suites were run twice at the final values, with identical results, to confirm determinism (test/test.js: 5 passing, test/test.cscal.js: 52 passing, test/test.ndarray.js: 73 passing, 0 failing; native suites skipped).
• make lint-javascript-tests TESTS_FILTER=".*/blas/base/cscal/.*" is clean.

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 unattended automated task. The test migration, the ULP measurements, and the minimality verification were performed by Claude Code; the approach mirrors the idiom established in previously merged migrations for this package family.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
blas/base/cscal	$\\color{green}379/379$ $\\color{green}+100.00\\%$	$\\color{green}17/17$ $\\color{green}+100.00\\%$	$\\color{green}4/4$ $\\color{green}+100.00\\%$	$\\color{green}379/379$ $\\color{green}+100.00\\%$

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