In Lexer's singleStringCharacter and doubleStringCharacter the lineContinuation is never parsed. The regression is isolated in "line-cont-string.js".