Function hasUTF8CodepointDifference

fuzz_stdlib_test.go:151–190

=========================================================================== Known differences helpers =========================================================================== hasUTF8CodepointDifference returns true if this pattern+input combination has known differences due to coregex matching by

(pattern, input string)

// This affects `.`, `\D`, `\W`, `\S`, negated character classes, and
// patterns that can match empty strings on multibyte input.
func hasUTF8CodepointDifference(pattern, input string) bool {
	// Check if input contains multibyte UTF-8 characters
	hasMultibyte := false
	for _, r := range input {
		if r >= 0x80 {
			hasMultibyte = true
			break
		}
	}
	if !hasMultibyte {
		return false
	}

	// Patterns that match at byte level vs codepoint level
	codepointPatterns := map[string]bool{
		`.`:         true, // dot matches any codepoint in stdlib, any byte in coregex
		`\D`:        true, // non-digit
		`\W`:        true, // non-word
		`\S`:        true, // non-space
		`[^a-z]`:    true, // negated class
		`[^0-9]`:    true, // negated class
		`[^a-zA-Z]`: true,
		// Empty-match patterns step by codepoint in stdlib, by byte in coregex
		``:    true, // empty pattern
		`a*`:  true, // can match empty
		`a?`:  true, // can match empty
		`a*?`: true, // can match empty
		`a??`: true, // can match empty
		`.*`:  true, // can match empty
		`.*?`: true, // can match empty
		`.?`:  true, // can match empty
		// Dot with captures also has codepoint differences
		`(.)`:  true,
		`(.)+`: true,
		`(.)*`: true,
		`(.)?`: true,
		`.+`:   true,
	}
	return codepointPatterns[pattern]
}

// isEmptyPatternCase returns true if this is an empty pattern case
// which has known differences in Split behavior.