refactor: align outliers in ndarray/iter with namespace error-construction conventions

[Planeshifter]

Aligning outliers in ndarray/iter with namespace majority patterns (random namespace pick, seed 20260426).

Description

This pull request normalizes a single drift item across six packages in the @stdlib/ndarray/iter namespace. Each outlier had exactly one throw new TypeError( '...' ) site that passed a plain string literal instead of routing the message through format(...) — diverging from the dominant pattern in the same file (4/5 = 80% of throws already used format(), including a static-string Error wrapped in format() for the read-only check) and from all three sibling packages with analogous dim-checks (subarrays, stacks, interleave-subarrays). Message text is preserved verbatim in every case.

Namespace summary

• Target namespace: ndarray/iter
• Members analyzed: 14 (all non-autogenerated)
• Random selection seed: 20260426
• Structural features analyzed: file tree, package.json shape, README section structure, test/benchmark/example file naming. All 14 packages structurally identical — no structural drift.
• Semantic features analyzed: public signature, return kind, validation prologue, error construction, JSDoc shape, @stdlib/* dependencies.
• Features with clear majority (≥75%): file tree (100%), package.json top-level keys (100%), README section sequence (100%, Usage → Notes → Examples → See Also), returnKind (100% iterator), jsdocShape.hasExample (100% true), errorConstruction (100% format at the package level when classified as "format dominates"). Within-file drift on a single throw site in 6/14 packages.
• Features without clear majority (excluded from drift detection): publicSignature, validationPrologue, dependencies — these legitimately vary by what each iterator does (shape vs x input, readonly vs order option, etc.).

Per-package notes

ndarray/iter/column-entries

Fixes a formatting inconsistency in ndarray/iter/column-entries where a single dim-check TypeError was thrown with a plain string literal rather than wrapped in format(), diverging from the other 4 of 5 throws in the file that already use format() — including a static-string Error on line 113. Across the ndarray/iter namespace, 8 of 14 sibling packages use format() exclusively, and all 3 analogous dim-check sites in subarrays, stacks, and interleave-subarrays wrap in format(). Message text is unchanged; this brings the package into conformance with the established pattern.

ndarray/iter/columns

Fixes a formatting inconsistency in ndarray/iter/columns where one TypeError throw was passing a plain string literal directly rather than routing through format(), diverging from the other 4 of 5 throws in the same file and from all 3 analogous dim-checks in sibling packages (subarrays, stacks, interleave-subarrays). Across the 14 ndarray/iter siblings, 8 use format() exclusively; the remaining 6 shared this same single-site deviation. Message text is unchanged.

ndarray/iter/matrices

Fixes a formatting inconsistency in ndarray/iter/matrices where the minimum-dimensions TypeError was constructed with a plain string literal rather than format(), diverging from the other four throws in the same file (80% conformance) and from all three sibling packages with analogous dimension checks (subarrays, stacks, interleave-subarrays). Across the fourteen ndarray/iter packages, eight already use format() exclusively; this was one of six exhibiting the same drift. Message text is unchanged.

ndarray/iter/matrix-entries

Fixes a formatting drift in ndarray/iter/matrix-entries where the minimum-dimensions TypeError was thrown with a plain string literal instead of format(), inconsistent with the other 4/5 throws in the same file and with all 3 analogous dim-checks across sibling packages (subarrays, stacks, interleave-subarrays). Across the broader ndarray/iter namespace, 8 of 14 siblings use format() exclusively; this package was among the 6 with this single plain-string outlier. The fix wraps the existing message text in format() without altering the message content.

ndarray/iter/row-entries

Fixes a formatting inconsistency in ndarray/iter/row-entries where the minimum-dimensions TypeError was constructed with a plain string literal rather than format(), making it the sole outlier among 5 throws in the file (the other 4, including a static-string Error, all use format()). Across the 14 ndarray/iter sibling packages, 8 use format() exclusively; all 3 siblings with analogous dimension-check throws (subarrays, stacks, interleave-subarrays) wrap them in format(). The fix wraps the throw in format() with the message text preserved verbatim.

ndarray/iter/rows

Fixes a formatting inconsistency in ndarray/iter/rows where the dimension-check TypeError was constructed with a plain string literal instead of format(), the pattern used by every other throw in the file (4 of the remaining 4 throws, including a static-string Error on line 113). Across the ndarray/iter namespace, 8 of 14 siblings use format() exclusively, and all 3 analogous dim-check sites in subarrays, stacks, and interleave-subarrays wrap their throws in format(). Message text is unchanged; this brings the one outlier throw in line with established convention.

Related Issues

This pull request has no related issues.

Questions

No.

Other

Validation

What was checked:

• Structural feature extraction across all 14 namespace members (file trees, package.json shape, README headings, test/benchmark/example file naming) — all identical.
• Per-package semantic feature extraction via 14 parallel sonnet agents (signature, validation prologue, error construction, JSDoc shape, dependencies).
• Three-agent drift validation: a semantic-review agent (opus), a cross-reference agent (opus), and a structural-review agent (sonnet). All three independently confirmed the drift as mechanical and safe.
• Direct smoke test: triggering the dim-check throw produces an identical error message to the pre-change version.

What was deliberately excluded:

• Features that legitimately vary across packages (publicSignature, validationPrologue, dependencies) — these reflect intentional semantic differences (e.g. entries/values/indices use options.order whereas the view-returning iterators use options.readonly, select-dimension takes a dim argument, interleave-subarrays takes arrays+ndims, etc.).
• No outlier was flagged as needing human judgment by the cross-reference agent; tests assert on TypeError constructor only and do not depend on exact message text (which is preserved regardless).

Checklist

• Read, understood, and followed the contributing guidelines.

AI Assistance

• Yes
• No

How AI assistance was used:

• 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 as part of an automated cross-package drift detection routine: a randomly selected stdlib namespace (ndarray/iter, seed 20260426) was scanned for majority-vs-outlier conventions across structural and semantic features, drift findings were independently validated by three review agents, and only mechanical, behavior-preserving corrections were applied. Each per-package commit was generated by Claude Code; a human maintainer should audit before promoting from draft.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
ndarray/iter/column-entries	$\color{green}290/290$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}290/290$ $\color{green}+100.00%$
ndarray/iter/columns	$\color{green}285/285$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}285/285$ $\color{green}+100.00%$
ndarray/iter/matrices	$\color{green}274/274$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}274/274$ $\color{green}+100.00%$
ndarray/iter/matrix-entries	$\color{green}291/291$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}291/291$ $\color{green}+100.00%$
ndarray/iter/row-entries	$\color{green}290/290$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}290/290$ $\color{green}+100.00%$
ndarray/iter/rows	$\color{green}285/285$ $\color{green}+100.00%$	$\color{green}27/27$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}285/285$ $\color{green}+100.00%$

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