Base64EncoderStream.java Source code

Java tutorial

Introduction

Here is the source code for Base64EncoderStream.java

Source

/*
    
   Licensed to the Apache Software Foundation (ASF) under one or more
   contributor license agreements.  See the NOTICE file distributed with
   this work for additional information regarding copyright ownership.
   The ASF licenses this file to You under the Apache License, Version 2.0
   (the "License"); you may not use this file except in compliance with
   the License.  You may obtain a copy of the License at
    
   http://www.apache.org/licenses/LICENSE-2.0
    
   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.
    
 */

import java.io.IOException;
import java.io.OutputStream;
import java.io.PrintStream;

/**
 * This class implements a Base64 Character encoder as specified in RFC1113.
 * Unlike some other encoding schemes there is nothing in this encoding
 * that indicates where a buffer starts or ends.
 *
 * This means that the encoded text will simply start with the first line
 * of encoded text and end with the last line of encoded text.
 *
 * @author <a href="mailto:deweese@apache.org">Thomas DeWeese</a>
 * @author <a href="mailto:vincent.hardy@eng.sun.com">Vincent Hardy</a>
 * @author      Chuck McManis
 * @version $Id: Base64EncoderStream.java 501495 2007-01-30 18:00:36Z dvholten $
 */
public class Base64EncoderStream extends OutputStream {

    /** This array maps the 6 bit values to their characters */
    private static final byte[] pem_array = {
            //   0   1   2   3   4   5   6   7
            'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', // 0
            'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', // 1
            'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', // 2
            'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f', // 3
            'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', // 4
            'o', 'p', 'q', 'r', 's', 't', 'u', 'v', // 5
            'w', 'x', 'y', 'z', '0', '1', '2', '3', // 6
            '4', '5', '6', '7', '8', '9', '+', '/' // 7
    };

    byte[] atom = new byte[3];
    int atomLen = 0;
    byte[] encodeBuf = new byte[4];
    int lineLen = 0;

    PrintStream out;
    boolean closeOutOnClose;

    public Base64EncoderStream(OutputStream out) {
        this.out = new PrintStream(out);
        closeOutOnClose = true;
    }

    public Base64EncoderStream(OutputStream out, boolean closeOutOnClose) {
        this.out = new PrintStream(out);
        this.closeOutOnClose = closeOutOnClose;
    }

    public void close() throws IOException {
        if (out != null) {
            encodeAtom();
            out.flush();
            if (closeOutOnClose)
                out.close();
            out = null;
        }
    }

    /**
     * This can't really flush out output since that may generate
     * '=' chars which would indicate the end of the stream.
     * Instead we flush out.  You can only be sure all output is
     * writen by closing this stream.
     */
    public void flush() throws IOException {
        out.flush();
    }

    public void write(int b) throws IOException {
        atom[atomLen++] = (byte) b;
        if (atomLen == 3)
            encodeAtom();
    }

    public void write(byte[] data) throws IOException {
        encodeFromArray(data, 0, data.length);
    }

    public void write(byte[] data, int off, int len) throws IOException {
        encodeFromArray(data, off, len);
    }

    /**
     * enocodeAtom - Take three bytes of input and encode it as 4
     * printable characters. Note that if the length in len is less
     * than three is encodes either one or two '=' signs to indicate
     * padding characters.
     */
    void encodeAtom() throws IOException {
        byte a, b, c;

        switch (atomLen) {
        case 0:
            return;
        case 1:
            a = atom[0];
            encodeBuf[0] = pem_array[((a >>> 2) & 0x3F)];
            encodeBuf[1] = pem_array[((a << 4) & 0x30)];
            encodeBuf[2] = encodeBuf[3] = '=';
            break;
        case 2:
            a = atom[0];
            b = atom[1];
            encodeBuf[0] = pem_array[((a >>> 2) & 0x3F)];
            encodeBuf[1] = pem_array[(((a << 4) & 0x30) | ((b >>> 4) & 0x0F))];
            encodeBuf[2] = pem_array[((b << 2) & 0x3C)];
            encodeBuf[3] = '=';
            break;
        default:
            a = atom[0];
            b = atom[1];
            c = atom[2];
            encodeBuf[0] = pem_array[((a >>> 2) & 0x3F)];
            encodeBuf[1] = pem_array[(((a << 4) & 0x30) | ((b >>> 4) & 0x0F))];
            encodeBuf[2] = pem_array[(((b << 2) & 0x3C) | ((c >>> 6) & 0x03))];
            encodeBuf[3] = pem_array[c & 0x3F];
        }
        if (lineLen == 64) {
            out.println();
            lineLen = 0;
        }
        out.write(encodeBuf);

        lineLen += 4;
        atomLen = 0;
    }

    /**
     * enocodeAtom - Take three bytes of input and encode it as 4
     * printable characters. Note that if the length in len is less
     * than three is encodes either one or two '=' signs to indicate
     * padding characters.
     */
    void encodeFromArray(byte[] data, int offset, int len) throws IOException {
        byte a, b, c;
        if (len == 0)
            return;

        // System.out.println("atomLen: " + atomLen +
        //                    " len: " + len +
        //                    " offset:  " + offset);

        if (atomLen != 0) {
            switch (atomLen) {
            case 1:
                atom[1] = data[offset++];
                len--;
                atomLen++;
                if (len == 0)
                    return;
                atom[2] = data[offset++];
                len--;
                atomLen++;
                break;
            case 2:
                atom[2] = data[offset++];
                len--;
                atomLen++;
                break;
            default:
            }
            encodeAtom();
        }

        while (len >= 3) {
            a = data[offset++];
            b = data[offset++];
            c = data[offset++];

            encodeBuf[0] = pem_array[((a >>> 2) & 0x3F)];
            encodeBuf[1] = pem_array[(((a << 4) & 0x30) | ((b >>> 4) & 0x0F))];
            encodeBuf[2] = pem_array[(((b << 2) & 0x3C) | ((c >>> 6) & 0x03))];
            encodeBuf[3] = pem_array[c & 0x3F];
            out.write(encodeBuf);

            lineLen += 4;
            if (lineLen == 64) {
                out.println();
                lineLen = 0;
            }

            len -= 3;
        }

        switch (len) {
        case 1:
            atom[0] = data[offset];
            break;
        case 2:
            atom[0] = data[offset];
            atom[1] = data[offset + 1];
            break;
        default:
        }
        atomLen = len;
    }

}