overwriting character arrays in C

scanf("%31s", userName);

This will limit userName to 31 characters. Your array is 32 chars long but you have to reserve 1 position for the null character indicating the end of the C-string.

There is no array bound control in C. You can "put" as many elements you want into a finite array.
But if you overide the array limit, you will write into other variables, or into the stack and you'll destroy it. This will cause a crash or unexpected behaviour.

Tron55555 wrote:
My instinct is to ask why C allows you to do this

By all means, go with your instinct. It is a far more interesting question. One of the problems with C is that, to do it well, you have to go really in depth with details of how to do low-level computing tasks as opposed to how to implement the logic of your program. And yet, so much software, even today, is still written in C. How do programmers do it, you ask? The answer - not well.

is there a reason that I should refrain from letting C overwrite the character array like this? Does it matter if my user (not that I have a user, but for argument's sake) enters something over 32 characters, since the array will hold it anyways?

As M. Imbert has already said, the array won't hold it. You just got lucky, or unlucky.

One of the problems that many self-taught programmers encounter (and usually refuse to accept) is that lots of really bad code works. Or, at least it works in the limited test cases that they try. Then they get very defensive about their code because "it works" on Linux, Windows, or in some other venue.

What you have to do in C is be very nit picky about null terminators, allocation sizes, result codes, etc. It all gets very complex in a hurry. Because of these details, C is good for learning how to program. Just don't expect to be very productive at it.

Analyze the following program and the output I've provided. Is it generating correct results?

#include <stdio.h>
#include <string.h>
main()
{
 char a[4];
 char b[32];
 char c[4];
 strcpy(a, "abc");
 strcpy(c, "qrs");
 strcpy(b, "123456789.123456789.123456789.12xyz");
 printf("addr of a: %p contents of a: %s ",a,a);
 printf("addr of b: %p contents of b: %s ",b,b);
 printf("addr of c: %p contents of c: %s ",c,c);
}

Here are the results. Are they correct?

addr of a: 0xbfffee8c contents of a: xyz
addr of b: 0xbfffee6c contents of b: 123456789.123456789.123456789.12xyz
addr of c: 0xbfffee68 contents of c: qrs

Now extrapolate to your scanf() question.

char userName[32];
scanf("%s", userName);
printf(" %s", userName);

I will also point out that my a, b, c variables are on the stack. Just beyond 'a' is the return address to 'main'. If I overwrite 'b' far enough, I'll overwrite the return address for 'main' which means I might return to some random location. This would apply to any function's return address.

NOTE: overwriting another variable by exceeding the size of an array, is also known as "Buffer Overflow" and is commonly used in "Security Exploits". You overwrite a string that is not bounds checked, and trash the return address, cause the program to run exploit code, and thus causing a program to perform activities it was not intended to do.