TCP Client Mathematical Operation not working properly

Question

The user enters 2 numbers and either +,-,/,* and should be getting an answer back. Problem is, it's not displaying what they entered properly and it's not calculating. Why is this? I have provided a screenshot to show what I am getting. How do I fix this? http://imageshack.us/photo/my-images/404/mathfn.jpg/

#include <iostream>
using namespace std;

#include <stdio.h>
#include <string.h>
#include <winsock.h>

// Function prototype
void StreamClient(char *szServer, short nPort);

// Helper macro for displaying errors
#define PRINTERROR(s)   \
        fprintf(stderr,"\n%s: %d\n", s, WSAGetLastError())

////////////////////////////////////////////////////////////

void main(int argc, char **argv)
{
    WORD wVersionRequested = MAKEWORD(1,1);
    WSADATA wsaData;
    int nRet;
    short nPort;

    //
    // Check for the host and port arguments
    //
    if (argc != 3)
    {
        fprintf(stderr,"\nSyntax: TCPTimeClient ServerName PortNumber\n");
        return;
    }

    nPort = atoi(argv[2]);

    //
    // Initialize WinSock and check the version
    //
    nRet = WSAStartup(wVersionRequested, &wsaData);
    if (wsaData.wVersion != wVersionRequested)
    {   
        fprintf(stderr,"\n Wrong version\n");
        return;
    }

    //
    // Go do all the stuff a datagram client does
    //
    StreamClient(argv[1], nPort);

    //
    // Release WinSock resources
    //
    WSACleanup();

    printf("\nPlease enter a floating point number: ");
float number1;
scanf("%f", &number1);

printf("\nPlease enter a mathematical operator (+,-,*,/): ");
char mathOperator;
scanf("%f", &mathOperator);


printf("\nPlease enter a floating point number: ");
float number2;
scanf("%f", &number2);

printf("\nCalculating %f %f %f...", &number1, &mathOperator, &number2);

printf("\nPlease enter a floating point number: ");
float number3;
scanf("%f", &number2);

}

////////////////////////////////////////////////////////////

void StreamClient(char *szServer, short nPort)
{
    int nRet;                       // return code
    char szBuf[256];                // client buffer area 
    char szSvr[256];                // server name

    LPHOSTENT lpHostEntry;          // host data structure
    SOCKET  theSocket;              // client socket
    SOCKADDR_IN saClient;           // socket address structure

    //
    // Get local machine name
    //
    nRet = gethostname(szSvr, sizeof(szSvr));

    //
    // Check for errors
    //
    if (nRet == SOCKET_ERROR)
    {
        PRINTERROR("gethostname()");
        return;
    }

    // 
    // Display an informational message
    //
    printf("Datagram Client [%s] sending to server [%s] on port %d...\n",
                                szSvr, szServer, nPort);

    //
    // Find the server
    //
    lpHostEntry = gethostbyname(szServer);
    if (lpHostEntry == NULL)
    {
        PRINTERROR("gethostbyname()");
        return;
    }

    //
    // Create a TCP/IP datagram socket
    //
    theSocket = socket(AF_INET,         // Address family
                       SOCK_STREAM,     // Socket type
                       0);              // Protocol

    //
    // Check for errors
    //
    if (theSocket == INVALID_SOCKET)
    {
        PRINTERROR("socket()");
        return;
    }

    //
    // Fill in the address structure of the server
    //
    saClient.sin_family = AF_INET;
    saClient.sin_addr = *((LPIN_ADDR)*lpHostEntry->h_addr_list);
                                        // ^ Client's address
    saClient.sin_port = htons(nPort);   // Port number from command line

    //
    // Connect to the server
    //
    nRet = connect(theSocket, 
            (struct sockaddr *)&saClient, 
            sizeof(saClient));

    //
    // Check for errors
    //
    if(nRet == SOCKET_ERROR)
    {
      PRINTERROR("Connect()");
      return;
    }

    //
    // Prepare some data to send to the server
    //
    sprintf(szBuf, "From the Client [%s]", szSvr);

    //
    // Send data to the server
    //
    nRet = send(theSocket,                  // Socket
                  szBuf,                    // Data buffer
                  (int)strlen(szBuf),       // Length of data
                  0);                       // Flags

    //
    // Check for errors
    //
    if (nRet == SOCKET_ERROR)
    {
        PRINTERROR("send()");
        closesocket(theSocket);
        return;
    }

    //
    // Zero out the incoming data buffer
    //
    memset(szBuf, 0, sizeof(szBuf));

    //
    // Wait for the reply
    //
    nRet = recv(theSocket,                  // Socket
                szBuf,                      // Receive buffer
                sizeof(szBuf),              // Length of receive buffer
                0);                         // Flags

    //
    // Check for errors
    //
    if (nRet == SOCKET_ERROR)
    {
        PRINTERROR("recv()");
        closesocket(theSocket);
        return;
    }

    //
    // Display the data that was received
    //
    printf("\n%s", szBuf);

    //
    // Close the socket
    //
    closesocket(theSocket);
    return;


}

Answer 1

I couldn't see why you were not able to input the char so I created a test program - sure enough same behavior! I don't scanf() alot, but I guessed that the scanf("%c") was grabbing whatever character was in the buffer - if any - and continuing on before we had a chance to enter anything. So I made it totally explicit - specifying 'stdin' for the fscanf/fgetchar stream and doing a fflush() to make sure it's empty before reading. Also, I stopped using scanf() at all just to get a single character. That was it! Here's the complete program:

printf("Please enter a floating point number: ");
float number1;
fscanf(stdin,"%f", &number1);

printf("Please enter a mathematical operator (+,-,*,/): ");
char mathOperator;
//fscanf(stdin,"%f", &mathOperator);

fflush(stdin);
mathOperator = (char)fgetc(stdin);

printf("Please enter a floating point number: ");
float number2;
fscanf(stdin,"%f", &number2);

printf("\nCalculating %f %c %f...", number1, mathOperator, number2);

float result = 0.0F;
switch (mathOperator) {
case '+':
    result = number1 + number2;
    break;
case '-':
    result = number1 - number2;
    break;
case '/':
    result = number1 / number2;
    break;
case '*':
    result = number1 * number2;
    break;
default:
    printf("\nInvalid operator entered!");
}

printf( "\n\nThe answer is %0.1f\n", result );

fflush(stdin);
char ch = getc(stdin);  // wait for keypress

And here's the output of a test run:

Please enter a floating point number: 3.4
Please enter a mathematical operator (+,-,*,/): *
Please enter a floating point number: 2

Calculating 3.400000 * 2.000000...

The answer is 6.8

[End Update]

The specific problem you're having with the printf() is that you're passing the addresses of the floating point values - not the values themselves.

printf("\nCalculating %f %f %f...", &number1, &mathOperator, &number2);

should be:

printf("\nCalculating %f %f %f...", number1, mathOperator, number2);

The next problem is 'mathOperator' - it looks like you're asking the user to input a char, but the data type specified int the printf() is float. That's not going to work.

So the printf really should be more like:

printf("\nCalculating %f %c %f...", number1, mathOperator, number2);

assuming 'mathOperator' has a '+', '-', etc. in it.

As far as why it's not calculating - I don't see (maybe it's just omitted for brevity) any code to actually do the calculation.

[Calculation Update]

Since the only allowed operators are '+', '-', '/', '*' and the operands are always floats, the calculation logic is pretty simple:

float result = 0.0F;
switch (mathOperator) {
case '+':
    result = number1 + number2;
    break;
case '-':
    result = number1 - number2;
    break;
case '/':
    result = number1 / number2;
    break;
case '*':
    result = number1 * number2;
    break;
}

Then you would probably want to add a printf() to output the result (and/or send it back via TCP):

printf( "The answer is" "%0.1f\n", result );

If you get into fancier calculations where you have to worry about operator precedence, type conversion, etc. a simple 'case' won't be enough. You'll want to look at parsers to build a parse tree from the expression and evaluate it according to some rules (grammar). But trust me - you want to master this first ;-)

See this question for an idea of what's involved.