Thread: Number user input

Ok here is my try at this. I think I understand what is going on pretty well. It seems to work and your right it's much smaller. I can dump the Validate function completely. I hope this is the right way to do this.

Code:

long int InputNext(int min, int max){

    char buffer[BUFSIZ];    
    char* pEnd;
    long int input;

    errno = 0;
   
    printf(CURSOR);

    while(errno != 42 ||
            errno == ERANGE ||
            *pEnd != '\0' ||
            input == 0 && buffer[0] != '0')
    {
       if ( fgets(buffer, sizeof buffer, stdin) != NULL ){
           buffer[strlen(buffer) -1] = '\0';
           input = strtol(buffer,&pEnd,10);            
       }
       printf(CURSOR);
    }
    return input;
}

If it works, so much the better, but people probably wont know what 42, the meaning of life, the universe, and everything, has to do with numerical input. (It actually corresponds to an illegal byte sequence in MinGW32, now that I look it up, but I can't say that's relevant. EINVAL is the only other err code besides ERANGE that matters.) Not to mention you de-reference pEnd before it points somewhere.

I don't know how long you've been coding, but you can stick to any line reading convention that you know. It's a lot easier to define the loop in terms of success as well.

Code:

while(fgets(buffer, sizeof buffer, stdin) != NULL)
{
   errno = 0;
   input = strtol(buffer, &pEnd, 10);
   if(*pEnd == '\n' && errno == 0) 
   {
      break;
   }
}

Not that you can't fix what you have.

LOL When I realized it returned 42 I chuckled a bit thinking Dennis Ritchie might have stuck a "Hitch-hiker's Guide to the Galaxy" reference in.

I am pretty green I started learning C on 05/01/2010 so only been at it a little while. I am going pretty slowly to insure I understand before moving on.

Not to mention you de-reference pEnd before it points somewhere.

Ooops! I was caught up in making the condition and didn't even notice that. I will need to be more careful and write out the algorithms first rather then building things trial and error as I have been doing.

It's a lot easier to define the loop in terms of success as well.

I had actually considered this first but went with the second guess. I should listen to my instincts more often.

Here is the fixed code. As I am going to use this for menus I need to add in the min and max values and not allow the user to just hit enter without entering something.

Code:

long int InputNext(int min, int max){

    char buffer[BUFSIZ];    
    char* pEnd;
    long int input;     
    
    printf(CURSOR);

    while (fgets(buffer, sizeof buffer, stdin) != NULL){
        errno = 0;
        printf("in loop\n");
        input = strtol(buffer,&pEnd,10);
        if ( *pEnd == '\n' && errno == 0
                           && input >= min
                           && input <= max
                           && buffer [0] != '\n'){
             break;
        }
        printf(CURSOR);
    }
    return input;
}

My question here is; in a complex conditional as I have above is this a good way to write this code style wise. To me lining up the &&'s makes it very readable.

I better get some sleep before commenting on your code.

I had to see if I could figure out how to do floating point number input from the above code and it was very simple. All I had to do is use strtod instead of strtol. They are a little different (only the base part). It's easier than I thought it would be. I just learned about arrays and pointer so using both was a bit intimidating. It took me a little bit to understand exactly what everything was doing. It is really cool how pEnd keep track of where it stopped.

here is the code for the floating point version.

Code:

double InputFloat(double min, double max){

    double input;
    char buffer[BUFSIZ];
    char* pEnd;

    printf(CURSOR);

    while (fgets(buffer, sizeof buffer, stdin) != NULL){
        errno = 0;
        input = strtod(buffer,&pEnd);
        if ( *pEnd == '\n' && errno == 0
                           && input >= min
                           && input <= max
                           && buffer[0] != '\n' ){
            break;
        }
        printf(CURSOR);
    }    
    return input;
}

if fgets() returns NULL(hit EOF), you are returning garbage.

Code:

double InputFloat(double min, double max){

    double input;
    char buffer[BUFSIZ];
    char* pEnd;
    input = min -1;

    printf(CURSOR);

    while (fgets(buffer, sizeof buffer, stdin) != NULL){
        errno = 0;
        input = strtod(buffer,&pEnd);
        if ( *pEnd == '\n' && errno == 0
                           && input >= min
                           && input <= max
                           && buffer[0] != '\n' ){
            break;
        }
        printf(CURSOR);
        input = min -1;
    }    
    return input;
}

I realize I am still way over my brain damaged head trying to get a stupid number from a user.

I will have to pick this apart piece by piece. I am determined to understand this.

You are making it unnecessarily complicated again.

Complexity seems to be my downfall.

My son asked what if the user wanted to know why they are being asked for input again instead of just a new cursor. So I added some error messages and a way to turn them on or off if desired.

Code:

/* 
 * File:   inputnumber.c
 * Author: jim
 *
 * Created on June 9, 2010, 11:09 AM
 */

/* *******************************************************
* Testbed for number input routines
*************************************************************/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>

#define MIN -100
#define MAX 100
#define EXTRA 1
#define VERBOSE_ON 1
#define VERBOSE_OFF 0

#define CURSOR ">>>"

long int InputNumber(int min, int max, int verbose);
void ClearBuf(void);

int main(void){

    long int number;


    printf("Enter a number (%d to %d)\n",MIN,MAX);
    number = InputNumber(MIN,MAX,VERBOSE_ON);

    printf("The number you entered = %ld\n",number);
    
    
    return 0;
}

long int InputNumber(int min, int max, int verbose){

    int i;
    char buffer[BUFSIZ];
    char* pEnd;
    long int input;


    do{
        // Making sure that buffer is empty when looped
        // not sure if I need to do this
        for (i = 0; i < sizeof buffer; i++){
            buffer[i] = 0;
        }

        // Make sure that all the values are 0 when looped
        errno = 0;
        input = 0;
        printf(CURSOR);

        if (fgets( buffer,sizeof buffer ,stdin)!= NULL){
            input = strtol(buffer,&pEnd,10);
            if ( buffer[0] != '\n' && *pEnd == '\n' || *pEnd == '\0'){
               // do something if successful stub
               // not necessary but could be useful in some circumstances
            }
            else {
                if ( input == 0 && buffer[0] == '\n'){
                    if (verbose == VERBOSE_ON) printf("You just hit return Please enter a number first!\n");
                }
                else if(input == 0 && *pEnd != '\n'){
                    if (verbose == VERBOSE_ON) printf("You entered letters not numbers!\n");
                }
                else
                    if (verbose == VERBOSE_ON) printf("You added %s",pEnd);

                // This is to make it loops if there is extra data
                // also when number to large reset input from junk
                if(errno != ERANGE) errno = EXTRA;
                if(errno == ERANGE) input = 0;
            }         
        }        
        if (errno == ERANGE)
            if (verbose == VERBOSE_ON) printf("The number is to large!\n",input);
        if (input < min || input > max && errno != ERANGE)
            if (verbose == VERBOSE_ON) printf("Your input %ld is out of range %d to %d\n",input,min,max);
        
    }
    while ( input < min || input > max ||errno == ERANGE || errno == EXTRA);
    
    return input;
}