Right -- that was my point. If they're using readline, they're subject to the rules of reading text (break the chunks into lines, stop at the 'end of file'). If they're using read, then they'll just get chunks of the file up to the file size.