In Java GB18030 charset, decode impl have this code

// Either 2 or 4 byte sequence follows
                        if ( sl - sp < 2 )
                            return CoderResult.UNDERFLOW;
                        byte2 = sa[sp + 1] & 0xFF;
                        inputSize = 2;

                        if (byte2 < 0x30)
                                return CoderResult.malformedForLength(1);

this code position is in GB18030.Decode#decodeArrayLoop, so when the char is double byte or four byte to show, I don’t know why the second byte(namely byte2) if malformed will throw malformed exception with 1 not 2.

Maybe it should be throw exception with 2?