more tweaks for parsing character strings * lex.ll (<SQ_STRING_START>\'\', <SQ_STRING_START>\', <SQ_STRING_START>[^\'\n\r]+): New patterns to replace <SQ_STRING_START>[^\'\n\r]*\'. (<SQ_STRING_START>[^\'\n\r]*\'): Delete. Attempt to correctly update input position for all patterns. (<DQ_STRING_START>\\{NL}): Only check for EOB or EOF if we are using the push lexer interface.