This tests the decoding of invalid UTF-8 sequences.
The following should be: "т<><D182>т <20><><EFBFBD>" or "т<><D182>т <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>":
т<EFBFBD><EFBFBD>т <20><><EFBFBD>