Some help for a Noobie!

Another Noobie,

Hi everyone. Im Barry, i just started using C over the last 4 months as part of my college course. I have a basic understanding of the language.

I have problems with some of the error trapping in my programming assignment and i was hoping you might be able to point me in the right direction.

Firstly:

Code:

---

void menu1(void)
{
void FileName (void) ;
int iChoice;
puts("\t\t\t\tMENU\n");
puts("\t\t\t1.\tEnter File Name");
puts("\t\t\t2.\tExit");
printf("Enter your choice: ");
scanf_s("%d", &iChoice);

switch (iChoice)
{
case 1:
        FileName();
case 2:
        exit(0);
default:
        puts("Please select either 1 or 2");
        menu1();
} /*end switch*/

} /*end menu1*/

---

I have the program compiling fine and looping when any number other than 1 or 2 are entered but when i enter a character it goes into an infinite loop.

This is probably pretty simple to fix but like i said im a noobie. Id appreciate any help you can give me!!

Thanks
Barry

It has to do with the switch statement jumping to a case based on an integer value, not a character value. You're assigning iChoice a character value (character object != integral object, an unsigned char is however). You might want to check the value of iChoice before continuing to the switch case.

Quote:

---

Originally Posted by barryr

Code:

---

scanf_s("%d", &iChoice);

---

---

I'm not familiar with scanf_s, but scanf does this when you use %d and enter a character. A possible solution might be to use getchar() and delete 48, to get the numerical value from the ascii code. When doing this you can also reject anything not in the range of 0-9.

ASCII:

Code:

---

48  0,    49  1,    50  2,    51  3,    52  4,    53  5,    54  6,    55  7,  56  8,    57  9

---

scanf_s is the "secure" version if you're using VS. Same goes for fopen, fwrite, etc. It is functionally the same as its "nonsecure" counter part.

Ok but does it share the infinite loop "feature" of it's scanf cousin?

It should (not that it's a good thing), like I said they are functionally the same, the secure versions usually checked for buffer overflows, null pointers, etc.

Quote:

---

Originally Posted by Subsonics

I'm not familiar with scanf_s, but scanf does this when you use %d and enter a character. A possible solution might be to use getchar() and delete 48, to get the numerical value from the ascii code. When doing this you can also reject anything not in the range of 0-9.

ASCII:

Code:

---

48  0,    49  1,    50  2,    51  3,    52  4,    53  5,    54  6,    55  7,  56  8,    57  9

---

---

When you say getchar() and delete 48 what do you mean exactly? I understand getchar but how do i delete 48?

Sorry for being clueless

Quote:

---

Originally Posted by barryr

When you say getchar() and delete 48 what do you mean exactly? I understand getchar but how do i delete 48?

Sorry for being clueless

---

If you use getchar, and enter 0 on stdin, that will correspond to ascii value 48. So, before you can use what getchar got for you as a number, subtract 48 as an offset.

The list I supplied is from ascii man page and shows the ascii codes for 0-9. So you could do something like this.

Code:

---

c = getchar();
c -= 48;

// Then you can use c as a number here.

---

You should not rely on its ascii representation nor use it anywhere in the program because it is difficult to understand what it means.
You can use is_digit to check if it is indeed a number. Then you can subtract by '0' to get the "real" number in integer format.
Example:
c = getchar();
c -= '0';
Do NOT subtract by 48. It's known as a magic number. A number thrown in somewhere which you might not have any idea of what it means. Therefore, subtract by '0'. That will give it a more meaningful description.

Quote:

---

Originally Posted by Elysia

You should not rely on its ascii representation nor use it anywhere in the program because it is difficult to understand what it means.
You can use is_digit to check if it is indeed a number. Then you can subtract by '0' to get the "real" number in integer format.
Example:
c = getchar();
c -= '0';
Do NOT subtract by 48. It's known as a magic number. A number thrown in somewhere which you might not have any idea of what it means. Therefore, subtract by '0'. That will give it a more meaningful description.

---

Agreed, that is a good advice. Or use a #define, it was just a quick example, it also does not check for values above 9 or below 0.

Yes, it does not. That might be the job of is_digit to make sure it really is a digit before subtracting.

Quote:

---

Originally Posted by Elysia

Yes, it does not. That might be the job of is_digit to make sure it really is a digit before subtracting.

---

But both getchar() and isdigit() would fail if the input was more than one byte.

But for this example it really isn't necessary to do any conversion at all and just use chars. The scanf could stay as well if it must, if only iChoice and %d is changed to char.

Code:

---

char iChoice;
scanf("%c", &iChoice);

switch(iChoice)
{
                case '1':
                {
                        puts("input was a one");
                        break;
                }
                case '2':
                {
                        puts("input was a two");
                        break;
                }
                default:
                {
                        puts("make a valid choice");
                }
}

---

Also involves a lot more code.

Yep, solving the problem by avoiding it. :-D

There are more elegant solutions...

Quote:

---

Originally Posted by Elysia

There are more elegant solutions...

---

Well, none of them is offered here.

Code:

---

char c = getchar();
if (is_digit(c))
{
    c -= '0';
    // Do whatever
}

---

For example.

Quote:

---

Originally Posted by Elysia

Code:

---

char c = getchar();
if (is_digit(c))
{
    c -= '0';
    // Do whatever
}

---

For example.

---

But the original poster, remember? Used a switch/case for a menu. And the above solution breaks in case of UTF16 or Unicode characters. I though you were on about a universal "bullet proof" input method for numbers. But in this case we don't need numbers, so why bother?

I suppose the OP was, yes. So in the OP's code, it would work fine, your example. For a generic solution of getting the actual number it represents (the -48 discussion), my example is better.
And neither handles UTF properly.

Quote:

---

Originally Posted by Elysia

For a generic solution of getting the actual number it represents (the -48 discussion), my example is better.

---

Yes, but IMO the -48 explains better what is going on, even though it is not a good practice to follow in code for an actual program.

Quote:

---

Originally Posted by Elysia

And neither handles UTF properly.

---

But what is the point of improving upon scanf's deficit, if there are still holes in the method.

I would argue that - '0' explains it even better (how can you subtract two letters from each other?).
To the last, then what is the point of using any method at all? Nothing is perfect. Remember that well. But there are ways to make it more perfect.

Wouldn't it be possible to make a univeral input method that recognices and handles any character encoding properly? I would think so, but I bet it's not going to be elegant. ;-)

Possible? No doubt. But messy, yes.
C++ handles these things much more smoothly.