Major Description
Tika's RTF parser processes the following rtf document incorrectly, applying the wrong character encoding to the parsed characters:
{\rtf1\ansi\ansicpg1252\fromtext \fbidis \deff0 {\fonttbl {\f0\fswiss\fcharset0 Arial;} {\f1\fswiss\fcharset204 Arial;}}
{\f1\fs20 \'d3\'e2\'e0\'e6\'e0\'e5\'ec\'fb\'e9 \'ea\'eb\'e8\'e5\'ed\'f2!\f0}\par
}
This document contains russian characters (\f1), but tika decodes these as latin due to the \f0 directive at the end of the group. The RTF parser should probably flush its pendingBytes buffer before processing directives such as these.