Class UnicodeEscapeOrError

java.lang.Object
org.ek9lang.compiler.phase1.UnicodeEscapeOrError
All Implemented Interfaces:
Consumer<EK9Parser.StringLitContext>
final class UnicodeEscapeOrError extends Object implements Consumer<EK9Parser.StringLitContext>
A backslash-u in a String literal must be followed by exactly four hex digits (E07903). The lexer only ever took backslash-u plus four ASCII hex digits as an escape, but the decoder re-read the text with parseInt(hex, 16), which also takes a sign and any Unicode digit: backslash-u "+041" became 'A' and backslash-u "-001" U+FFFF (front-end audit FE73). Steve: a new code, so a malformed escape is named for what it is. Checked in a double-quoted String and in the text of a backtick String; not in a directive's String, which is test metadata.

The same walk reports a surrogate escape (D800-DFFF) that is not a high one directly followed by a low one (E07904, Steve 2026-11-04): a String or Character is a sequence of Unicode code points, and half of a code point above U+FFFF is not one. A pair of escapes in a String is one character; in a Character literal a surrogate escape is always alone, so always an error.

(No backslash-u appears literally in this comment: javac decodes unicode escapes even in comments.)