Compare FoldCase in Regexp.equals/hashCode for LITERAL/CHAR_CLASS.

[jemmix]

Fixes #211.

Regexp.equals() compared only the runes for LITERAL/CHAR_CLASS, so alternation factoring (Parser.factor round 1) merged a folded literal with its case-sensitive twin and the fold leaked into the merged arm: (?i:Z)x|Z matched lowercase "z" (the case-sensitive arm was effectively deleted). Go's regexp/syntax compares Flags&FoldCase here; the port dropped it.

This adds FoldCase to both equals() and hashCode() (factoring groups via hashCode first, so the hash must agree with equality).

Adds FoldCaseFactoringTest: the fold no longer leaks ((?i:Z)x|Z, (?i:[Z])x|[Z] reject "z"), and both arms still match ("Zx" via the folded arm, "Z" via the case-sensitive one).

Verified locally with ./gradlew check on JDK 8: 1831 tests, 0 failures, verifyGoogleJavaFormat + license green.

Disclosure (as in #211): found while fuzzing https://github.com/jemmix/tdfa-jvm against re2j; patch and tests composed agentically with GLM 5.3 assistance; verification run locally by me.