In bug 5005831, the following new methods were added:

  public byte[] getBytes(byte [], Charset);
  public String(byte bytes[], Charset charset);
  public String(byte bytes[], int offset, int length, Charset charset);

In all cases, we did not specify the behaviour on invalid input. The Charset parameter dictates what that behaviour should be so we should specify it.

Replace this sentence in all of the new methods:

   * <p> The behavior of this method when the string cannot be encoded in the
   * given charset is unspecified.

With this one:

   * <p> This method always replaces malformed-input and unmappable-character
   * sequences with this charset's default replacement XXX.

where XXX = "string" in String(byte [], Charset), String(byte[], int, int, Charset)
            "byte array" in String.getBytes(Charset)