While digging into ReplaceText today, I noticed that org.apache.nifi.processors.standard.util.NLKBufferedReader is actually a direct copy of the JDK version of BufferedReader, with a few minor code edits made on top of it so that new line characters are kept. This code was in the original code commit from way back.
It feels like there is a licensing issue here, as this is Oracle source code. Joe? --Peter