test: migrate math/base/tools/sum-series to ULP-based assertions

[kgryte]

Resolves a part of #11352.

Description

What is the purpose of this pull request?

This pull request:

• migrates the tests for math/base/tools/sum-series from tolerance-based assertions (abs( actual - expected ) < EPS) to ULP difference testing using @stdlib/assert/is-almost-same-value.
• removes the now unused @stdlib/math/base/special/abs and @stdlib/constants/float64/eps requires from the affected test files.
• normalizes the affected assertion messages to 'returns expected value', per the migration guidance in [RFC]: Migrate math/base/special packages from relative tolerance testing to ULP difference testing (tracking issue) #11352.

Files changed

• test/test.js (4 assertions)
• test/es2015-generator/index.js (1 assertion)

No non-test files are touched, and no implementation behavior changes.

ULP bounds

Each assertion uses the minimum required ULP value, measured against the values actually produced in this environment:

Test file	Assertion	Measured ULP difference	ULP bound used
test/test.js	sum of series from a closure	3	3
test/test.js	sum of series from a closure (generators unsupported)	3	3
test/test.js	sum of series with a specified initial value	0 (bit-exact)	0
test/test.js	sum of series with a specified initial value (generators unsupported)	0 (bit-exact)	0
test/es2015-generator/index.js	sum of series from a generator	3	3

The two assertions comparing against integer-valued results (maxTerms: 3, expected 6) were already exact t.strictEqual comparisons and are intentionally left unchanged.

Tightness was verified by lowering the bound: at 2 ULP, exactly the three 3-ULP assertions fail, so 3 is minimal. The 0-ULP assertions are bit-exact (isSameValue), so they cannot be tightened further. The suite was run twice at the final bounds with identical results (9/9 passing), so the measured differences are not FMA/arch-dependent flake in this environment.

make test TESTS_FILTER=".*/math/base/tools/sum-series/.*"
# total: 9, passing: 9

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?

Two points worth a reviewer's eye:

1. Scope. test/es2015-generator/index.js is converted alongside test/test.js, since test.js requires it and it carried the same tolerance idiom. This mirrors the sibling migration for math/base/tools/continued-fraction (test: migrate math/base/tools/continued-fraction to ULP-based assertions #15573), which converted both files together. Happy to narrow the PR to test/test.js if you would prefer the generator test handled separately.
2. Lint directive. The eslint-disable directive in test/es2015-generator/index.js is updated from no-restricted-syntax, no-plusplus, require-jsdoc to no-restricted-syntax, no-plusplus, n/no-unsupported-features/es-syntax. This is a pre-existing lint failure in that file, not one introduced here, but since the PR touches the file it would otherwise ship red. The same rule was added to the directive in the continued-fraction migration, whose generator test now lints clean. Let me know if you would rather this be split into its own change.

Other

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

• The original assertions used an absolute EPS tolerance rather than the relative EPS * abs( expected ) form shown in [RFC]: Migrate math/base/special packages from relative tolerance testing to ULP difference testing (tracking issue) #11352. At the magnitude of these results (log1p( 0.5 ) ≈ 0.405), one ULP is ≈ 5.55e-17, so the old absolute EPS bound (≈ 2.22e-16) corresponded to roughly 4 ULP. The new 3 ULP bound is therefore slightly tighter than what the tests previously accepted.
• Linting is clean for both changed files under etc/eslint/.eslintrc.tests.js.

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 scheduled task. The agent selected the package, studied the prior math/base/tools/continued-fraction migration to match the project's settled idiom, measured the per-assertion ULP differences, tightened each bound to its minimum, and verified tests and linting locally. Opened as a draft for human review.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
math/base/tools/sum-series	$\\color{red}219/246$ $\\color{green}+89.02\\%$	$\\color{red}14/15$ $\\color{green}+93.33\\%$	$\\color{red}1/2$ $\\color{green}+50.00\\%$	$\\color{red}219/246$ $\\color{green}+89.02\\%$

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