feat: add eval for UTF-16 character predicate streams

[martinfrancois]

Problem

The Streams capture workflow missed a character loop that can become String.chars().allMatch(...). This is eval-worthy because a superficially cleaner rewrite can change two contracts: allMatch is vacuously true for empty input, and codePoints() changes UTF-16 code-unit iteration into Unicode code-point iteration.

Area

Evals or scoring

Describe the solution you want

Add a reference eval for replacing a non-empty universal charAt predicate loop with String.chars().allMatch(...) while preserving empty-input and supplementary-character behavior.

Code before the prompt was executed

boolean isNumeric(String value) {
    if (value.isEmpty()) {
        return false;
    }
    for (int i = 0; i < value.length(); i++) {
        if (!Character.isDigit(value.charAt(i))) {
            return false;
        }
    }
    return true;
}

Prompt that caused the implementation

The original prompt required a complete Java-source API-reuse audit:

• Audit every tracked Java source for worthwhile replacements with the Java standard library or a directly declared dependency.
• Implement only changes that clearly reduce code, complexity, duplication, or maintenance risk.
• Preserve observable behavior, ordering, failure semantics, null handling, concurrency, security, and relevant performance characteristics.
• Add focused regression tests for each affected boundary.

The prompt should have activated the Streams skill for a pure universal predicate loop.

Later prompt that exposed the issue

A maintainer later asked whether the Streams capture issue had been created and requested that the capture workflow cover cases like this.

Prompt-produced code before maintainer correction

boolean isNumeric(String value) {
    return !value.isEmpty() && value.chars().allMatch(Character::isDigit);
}

The implementation is preferred, but its important equivalence proof was not captured.

Why the prompt-produced code is weak

The code is not weak; the capture workflow is. Existing predicate-loop guidance explicitly excludes character scans, so it cannot teach when a character scan is safely result-producing. Without a focused eval, an agent may drop the empty guard or choose codePoints() because it appears more Unicode-aware, silently changing behavior.

Behavior-equivalence analysis

allMatch returns true for an empty stream, so !value.isEmpty() is required. charAt and String.chars() both inspect UTF-16 code units. A supplementary mathematical digit such as 𝟙 is two surrogate code units and remains rejected; codePoints() would see one digit and accept it. Tests must cover ASCII digits, BMP Arabic-Indic digits, empty and mixed strings, signs/whitespace, and a supplementary digit.

Maintainer-preferred code

boolean isNumeric(String value) {
    return !value.isEmpty() && value.chars().allMatch(Character::isDigit);
}

Why the replacement is better

The terminal directly expresses the universal predicate and preserves short-circuiting, while the guard and chars() choice make the old contracts explicit.

Desired eval behavior

• Reward activation during the original audit prompt.
• Reward preserving the non-empty guard.
• Reward matching the original iteration unit: charAt to chars().
• Reward explaining vacuous truth and the UTF-16/code-point distinction.
• Reward focused supplementary-character tests.

Anti-patterns the eval should reject

• Dropping the empty guard.
• Replacing chars() with codePoints() without an explicit contract change.
• Boxing the primitive stream.
• Replacing the loop with a regex without proving identical Unicode semantics.

Suggested eval name

utf16-character-allmatch

Examples

Use the helper above as the self-contained reference task. Add Java 8 compatibility entries for String.chars() and String.codePoints() if runtime guidance is added.

Alternatives considered

Appending this to the general predicate-loop issue #49 was rejected because #49 deliberately excludes character scans and does not cover iteration-unit semantics.

Current workaround

Manually compare charAt, chars, and codePoints, and add explicit empty/supplementary test cases.

Additional context

No user-facing contract change is intended.

AI Assistance

• AI-assisted issue
• I confirm I understand and reviewed this request

AI prompts / session logs (optional)

Prepared with Codex assistance from a maintainer review of a behavior-preserving Java API-reuse audit.