jdk/src/share/classes/sun/io/ByteToCharEUC2.java
author ohair
Tue, 25 May 2010 15:58:33 -0700
changeset 5506 202f599c92aa
parent 2921 d9d491a5a169
permissions -rw-r--r--
6943119: Rebrand source copyright notices Reviewed-by: darcy, weijun
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
2921
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     1
/*
5506
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
     2
 * Copyright (c) 1997, Oracle and/or its affiliates. All rights reserved.
2921
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     3
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     4
 *
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     5
 * This code is free software; you can redistribute it and/or modify it
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     6
 * under the terms of the GNU General Public License version 2 only, as
5506
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
     7
 * published by the Free Software Foundation.  Oracle designates this
2921
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
     8
 * particular file as subject to the "Classpath" exception as provided
5506
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
     9
 * by Oracle in the LICENSE file that accompanied this code.
2921
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    10
 *
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    11
 * This code is distributed in the hope that it will be useful, but WITHOUT
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    12
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    13
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    14
 * version 2 for more details (a copy is included in the LICENSE file that
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    15
 * accompanied this code).
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    16
 *
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    17
 * You should have received a copy of the GNU General Public License version
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    18
 * 2 along with this work; if not, write to the Free Software Foundation,
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    19
 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    20
 *
5506
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
    21
 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
    22
 * or visit www.oracle.com if you need additional information or have any
202f599c92aa 6943119: Rebrand source copyright notices
ohair
parents: 2921
diff changeset
    23
 * questions.
2921
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    24
 */
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    25
package sun.io;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    26
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    27
import sun.nio.cs.ext.DoubleByte;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    28
import static sun.nio.cs.CharsetMapping.*;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    29
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    30
public abstract class ByteToCharEUC2 extends ByteToCharConverter
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    31
{
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    32
    private final int G0 = 0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    33
    private final int G1 = 1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    34
    private final int SS2 =  0x8E;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    35
    private final int SS3 =  0x8F;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    36
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    37
    private int firstByte, state;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    38
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    39
    private DoubleByte.Decoder dec;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    40
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    41
    public ByteToCharEUC2(DoubleByte.Decoder dec) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    42
        super();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    43
        state = G0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    44
        this.dec = dec;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    45
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    46
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    47
    char decodeSingle(int b) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    48
        return dec.decodeSingle(b);
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    49
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    50
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    51
    char decodeDouble(int b1, int b2) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    52
        return dec.decodeDouble(b1, b2);
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    53
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    54
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    55
    /**
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    56
      * flush out any residual data and reset the buffer state
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    57
      */
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    58
    public int flush(char[] output, int outStart, int outEnd)
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    59
       throws MalformedInputException
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    60
    {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    61
       if (state != G0) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    62
          reset();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    63
          badInputLength = 0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    64
          throw new MalformedInputException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    65
       }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    66
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    67
       reset();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    68
       return 0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    69
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    70
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    71
    /**
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    72
     *  Resets the converter.
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    73
     */
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    74
    public void reset() {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    75
       state = G0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    76
       charOff = byteOff = 0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    77
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    78
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    79
    /**
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    80
     * Character conversion
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    81
     */
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    82
    public int convert(byte[] input, int inOff, int inEnd,
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    83
                       char[] output, int outOff, int outEnd)
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    84
        throws UnknownCharacterException, MalformedInputException,
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    85
               ConversionBufferFullException
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    86
    {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    87
        int       byte1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    88
        char      outputChar = UNMAPPABLE_DECODING;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    89
        byteOff = inOff;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    90
        charOff = outOff;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    91
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    92
        while (byteOff < inEnd) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    93
            byte1 = input[byteOff] & 0xff;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    94
            switch (state) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    95
            case G0:
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    96
                if (byte1 == SS2 ||                // no general support
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    97
                    byte1 == SS3 ) {               //    for g2 or g3
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    98
                    badInputLength = 1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
    99
                    throw new MalformedInputException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   100
                }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   101
                if ( byte1 <= 0x9f )               // < 0x9f has its own table
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   102
                    outputChar = decodeSingle(byte1);
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   103
                else
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   104
                    if (byte1 < 0xa1 || byte1 > 0xfe) {  // byte within range?
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   105
                        badInputLength = 1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   106
                        throw new MalformedInputException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   107
                    } else {                       // G1 set first byte
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   108
                        firstByte = byte1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   109
                        state = G1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   110
                    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   111
                break;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   112
            case G1:
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   113
                state = G0;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   114
                if ( byte1 < 0xa1 || byte1 > 0xfe) {  // valid G1 set second byte
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   115
                    badInputLength = 1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   116
                    throw new MalformedInputException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   117
                }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   118
                outputChar = decodeDouble(firstByte, byte1);
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   119
                break;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   120
            }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   121
            if (state == G0) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   122
                if (outputChar == UNMAPPABLE_DECODING) {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   123
                    if (subMode)
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   124
                        outputChar = subChars[0];
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   125
                    else {
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   126
                        badInputLength = 1;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   127
                        throw new UnknownCharacterException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   128
                    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   129
                }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   130
                if (charOff >= outEnd)
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   131
                    throw new ConversionBufferFullException();
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   132
                output[charOff++] = outputChar;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   133
            }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   134
            byteOff++;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   135
         }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   136
         return charOff - outOff;
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   137
    }
d9d491a5a169 6843578: Re-implement IBM doublebyte charsets
sherman
parents:
diff changeset
   138
}