1

私には学校の課題があります。jpg画像を送信し、100バイトのグループに分割し、破損し、CRCチェックを使用してエラーを特定し、最終的に元の形式に戻るまで再送信します。

実質的には準備ができていますが、新しい画像をチェックアウトすると、エラーが表示されます。

誰かが以下の私のコードを見て、問題が何であるか理解できないので、この論理的な間違いを見つけることができれば本当にありがたいです:S

写真やエラーパターンを含む必要なすべてのデータを含むファイルについては、次のリンクからダウンロードできます:http://rapidshare.com/#!download|932tl2|443122762|Data.zip|739

画像ファイルとエラーファイルのコードのパスを変更することを忘れないでください。

package networks;

import java.io.*;               // for file reader
import java.util.zip.CRC32;     // CRC32 IEEE (Ethernet)

public class Main {

/**
 * Reads a whole file into an array of bytes.
 * @param file The file in question.
 * @return Array of bytes containing file data.
 * @throws IOException Message contains why it failed.
 */
public static byte[] readFileArray(File file) throws IOException {
    InputStream is = new FileInputStream(file);
    byte[] data=new byte[(int)file.length()];
    is.read(data);
    is.close();
    return data;
}

/**
 * Writes (or overwrites if exists) a file with data from an array of bytes.
 * @param file The file in question.
 * @param data Array of bytes containing the new file data.
 * @throws IOException Message contains why it failed.
 */
public static void writeFileArray(File file, byte[] data) throws IOException {
    OutputStream os = new FileOutputStream(file,false);
    os.write(data);
    os.close();
}

/**
 * Converts a long value to an array of bytes.
 * @param data The target variable.
 * @return Byte array conversion of data.
 * @see http://www.daniweb.com/code/snippet216874.html
 */
public static byte[] toByta(long data) {
    return new byte[] {
        (byte)((data >> 56) & 0xff),
        (byte)((data >> 48) & 0xff),
        (byte)((data >> 40) & 0xff),
        (byte)((data >> 32) & 0xff),
        (byte)((data >> 24) & 0xff),
        (byte)((data >> 16) & 0xff),
        (byte)((data >> 8) & 0xff),
        (byte)((data >> 0) & 0xff),
    };
}

/**
 * Converts a an array of bytes to long value.
 * @param data The target variable.
 * @return Long value conversion of data.
 * @see http://www.daniweb.com/code/snippet216874.html
 */
public static long toLong(byte[] data) {
    if (data == null || data.length != 8) return 0x0;
    return (long)(
        // (Below) convert to longs before shift because digits
        // are lost with ints beyond the 32-bit limit
        (long)(0xff & data[0]) << 56 |
        (long)(0xff & data[1]) << 48 |
        (long)(0xff & data[2]) << 40 |
        (long)(0xff & data[3]) << 32 |
        (long)(0xff & data[4]) << 24 |
        (long)(0xff & data[5]) << 16 |
        (long)(0xff & data[6]) << 8  |
        (long)(0xff & data[7]) << 0
    );
}

public static byte[] nextNoise(){
    byte[] result=new byte[100];
    // copy a frame's worth of data (or remaining data if it is less than frame length)
    int read=Math.min(err_data.length-err_pstn, 100);
    System.arraycopy(err_data, err_pstn, result, 0, read);
    // if read data is less than frame length, reset position and add remaining data
    if(read<100){
        err_pstn=100-read;
        System.arraycopy(err_data, 0, result, read, err_pstn);
    }else // otherwise, increase position
        err_pstn+=100;
    // return noise segment
    return result;
}

/**
 * Given some original data, it is purposefully corrupted according to a
 * second data array (which is read from a file). In pseudocode:
 * corrupt = original xor corruptor
 * @param data The original data.
 * @return The new (corrupted) data.
 */
public static byte[] corruptData(byte[] data){
    // get the next noise sequence
    byte[] noise = nextNoise();
    // finally, xor data with noise and return result
    for(int i=0; i<100; i++)data[i]^=noise[i];
    return data;
}

/**
 * Given an array of data, a packet is created. In pseudocode:
 * frame = corrupt(data) + crc(data)
 * @param data The original frame data.
 * @return The resulting frame data.
 */
public static byte[] buildFrame(byte[] data){
    // pack = [data]+crc32([data])
    byte[] hash = new byte[8];
    // calculate crc32 of data and copy it to byte array
    CRC32 crc = new CRC32();
    crc.update(data);
    hash=toByta(crc.getValue());
    // create a byte array holding the final packet
    byte[] pack = new byte[data.length+hash.length];
    // create the corrupted data
    byte[] crpt = new byte[data.length];
    crpt = corruptData(data);
    // copy corrupted data into pack
    System.arraycopy(crpt, 0, pack, 0, crpt.length);
    // copy hash into pack
    System.arraycopy(hash, 0, pack, data.length, hash.length);
    // return pack
    return pack;
}

/**
 * Verifies frame contents.
 * @param frame The frame data (data+crc32).
 * @return True if frame is valid, false otherwise.
 */
public static boolean verifyFrame(byte[] frame){
    // allocate hash and data variables
    byte[] hash=new byte[8];
    byte[] data=new byte[frame.length-hash.length];
    // read frame into hash and data variables
    System.arraycopy(frame, frame.length-hash.length, hash, 0, hash.length);
    System.arraycopy(frame, 0, data, 0, frame.length-hash.length);
    // get crc32 of data
    CRC32 crc = new CRC32();
    crc.update(data);
    // compare crc32 of data with crc32 of frame
    return crc.getValue()==toLong(hash);
}

/**
 * Transfers a file through a channel in frames and reconstructs it into a new file.
 * @param jpg_file File name of target file to transfer.
 * @param err_file The channel noise file used to simulate corruption.
 * @param out_file The name of the newly-created file.
 * @throws IOException
 */
public static void transferFile(String jpg_file, String err_file, String out_file) throws IOException {
    // read file data into global variables
    jpg_data = readFileArray(new File(jpg_file));
    err_data = readFileArray(new File(err_file));
    err_pstn = 0;
    // variable that will hold the final (transfered) data
    byte[] out_data = new byte[jpg_data.length];
    // holds the current frame data
    byte[] frame_orig = new byte[100];
    byte[] frame_sent = new byte[100];
    // send file in chunks (frames) of 100 bytes
    for(int i=0; i<Math.ceil(jpg_data.length/100); i++){
        // copy jpg data into frame and init first-time switch
        System.arraycopy(jpg_data, i*100, frame_orig, 0, 100);
        boolean not_first=false;
        System.out.print("Packet #"+i+": ");
        // repeat getting same frame until frame crc matches with frame content
        do {
            if(not_first)System.out.print("F");
            frame_sent=buildFrame(frame_orig);
            not_first=true;
        }while(!verifyFrame(frame_sent));   // usually, you'd constrain this by time to prevent infinite loops (in
                                            // case the channel is so wacked up it doesn't get a single packet right)
        // copy frame to image file
        System.out.println("S");
        System.arraycopy(frame_sent, 0, out_data, i*100, 100);
    }
    System.out.println("\nDone.");
    writeFileArray(new File(out_file),out_data);
}

// global variables for file data and pointer
public static byte[] jpg_data;
public static byte[] err_data;
public static int    err_pstn=0;

public static void main(String[] args) throws IOException {

    // list of jpg files
    String[] jpg_file={
        "C:\\Users\\Stefan\\Desktop\\Data\\Images\\photo1.jpg",
        "C:\\Users\\Stefan\\Desktop\\Data\\Images\\photo2.jpg",
        "C:\\Users\\Stefan\\Desktop\\Data\\Images\\photo3.jpg",
        "C:\\Users\\Stefan\\Desktop\\Data\\Images\\photo4.jpg"
    };

    // list of error patterns
    String[] err_file={
        "C:\\Users\\Stefan\\Desktop\\Data\\Error Pattern\\Error Pattern 1.DAT",
        "C:\\Users\\Stefan\\Desktop\\Data\\Error Pattern\\Error Pattern 2.DAT",
        "C:\\Users\\Stefan\\Desktop\\Data\\Error Pattern\\Error Pattern 3.DAT",
        "C:\\Users\\Stefan\\Desktop\\Data\\Error Pattern\\Error Pattern 4.DAT"
    };
    // loop through all jpg/channel combinations and run tests
    for(int x=0; x<jpg_file.length; x++){
        for(int y=0; y<err_file.length; y++){
            System.out.println("Transfering photo"+(x+1)+".jpg using Pattern "+(y+1)+"...");
            transferFile(jpg_file[x],err_file[y],jpg_file[x].replace("photo","CH#"+y+"_photo"));
        }
    }

}

}

4

1 に答える 1

2

Learning how to debug is probably one of the most important lessons for development, so, as I would end up following these steps I will describe how you can approach it.

  1. Make a unit test for your program.
  2. You have it hardcoded for 100 byte chunks. Make that variable, so your unit test can change the value, to perhaps 5 or 10 at first.
  3. Your first test is to take a known 5 byte chunk.
  4. It will be corrupted, verify that the corruption is what was expected. If not, fix it.
  5. Then test transmitting, verify that what was transmitted is what was received.
  6. Then test the fixing.
  7. Then do it with a larger size, perhaps 20, then do 100, then an image.

Having the size hard-coded is bad practice. Writing code without unit tests is bad, as you may (will) then write code that is not testable, so is harder to debug.

This will enable you to find which functions are not working according to your expectations, and you can determine what is going on, you may need to refactor your code and break functions up into smaller units.

于 2011-01-17T21:51:52.193 に答える