Java Socket server with C client problem

[rnvrnv]

Hi,
I am new to java socket specially if the client is in C.

The java server gets the connection and gets something from the C client but I am it is not reading the message correctly. the C client sends the data in byte stream which i try to get as follows:

BufferedInputStream bis = new BufferedInputStream(clientSocket
						.getInputStream());
DataInputStream in =new DataInputStream (bis);
byte[] b = new byte[200];

in.read(b);
for (int i = 0; i < b.length; i++) {
   clientmsg+=(int)b[i];
}
System.out.println("clientmsg: "+clientmsg);

the problem is i am not able to decode the message. the structure of the message from the client is as follows:

typedef struct user
{
    u_int 32 username
    u_int 32 city
    u_int 8  validuser /* 1-yes 0-no */
} user _t, * user _p

hope someone can help me.
regards,
rnv

[Norm]

Can you define the bytes being sent to the Java program byte by byte vs showing C code?
Have you tried debugging the code by printing out the bytes as they are read so you can see what is being received?
What is the type of clientmsg?
Have you tried using the String() constructor to build the String vs the concatenate technique you are using.

[rnvrnv]

Sorry I do not have the C code.

I read something like this:
000000048110079120521593741290122664

But I guess it is wrong. my original post contains how I get this thing but I think I am doing something wrong.

actually I should be getting the following structure:

typedef struct user
{
    u_int 32 userid 
    u_int 32 cityid
    u_int 8  validuser /* 1-yes 0-no */
} user _t, * user _p

the values could be
123 //user id
987 //city id
1 // successfull

is it a problem of encoding between java (windows) and C (linux) system?

regards,
rnv

[Norm]

I read something like this:

What does 000000048110079120521593741290122664 represent?
Is it hex? 00 00 00 04 81 10 07 91 20 52 15 93 74 12 90 12 26 64 (18 bytes long)

Sorry I do not have the C code

What is this? I thought it looked like C code:

typedef struct user
{
    u_int 32 userid 
    u_int 32 cityid
    u_int 8  validuser /* 1-yes 0-no */
} user _t, * user _p

is it a problem of encoding between java (windows) and C (linux) system?

Depends. A byte in java should be the same byte in C. Anything else could be different.

[rnvrnv]

Here is how I would like to read message from the C socket client in my Java Socket server. the example shows how to read 4 byte followed by 1 byte int.

to get int from 4byte array I used getIntFromByteArray method. to get 1byte value i simply get the byte.

Is it the correct way?

BufferedInputStream bis = new BufferedInputStream(clientSocket.getInputStream());
DataInputStream in = new DataInputStream(bis);

// read 4 byte user id
byte[] useridByte = new byte[4];
in.read(useridByte); // Bytes are read into

//read 1 byte age
byte[] ageByte= new byte[1];
in.read(versionbyteByte);

// get userid
int userid = getIntFromByteArray(useridByte);
// get age
byte age = ageByte[0];

public static int getIntFromByteArray(byte[] b) {
  int number = 0;
  for (int i = 0; i < 4; ++i) {
     number |= (b[3 - i] & 0xff) << (i << 3);
  }
  return number;
}

[Norm]

Use the same technique I showed in the other thread. Put the bytes into a ByteArrayInputStream and have your Data stream wrap that. Then run your code and print out the results and compare it the what you put into the BAIS.

[spygas]

Does your C client convert struct to network-byte-order before send()? It's requiered if you are sending structs containing fields larger than char (like int, long, etc). Check this out: htons(), htonl(), ntohs(), ntohl()
Btw, I would recommend you to convert byte array to int in more simple way:

public static int getIntFromByteArray(byte[] b) {
    int number = 0;
    for (int x = 0; x < b.length; x++)
        number = (number << 8) | b[x];
    return number;
}