why DOS prompt crashes when I run this compiled code?

Hi

The computer I'm using has Win Visa installed. I'm using Code::Blocks and run compiled from within the Code::Blocks.

CODE 1 compiles and runs fine. CODE 2 compiles but when I try to run it the command prompt (DOS window) crashes without even running the compiled code. I have even tried Dev-C++ the same problem is encountered. At the bottom you can find the problem details shown by the crashed DOS window and also a linked screenshot of crashed window. Please help me to solve this issue.

CODE 1

Code:

// calculating area of a circle using user-defined function

#include <iostream>
#include <cstdlib>

using namespace std;

float area(float dummy);

int main()

{
    float r; float a;

	cout << "enter radius: "; cin >> r;

	a = area(r);

	cout << a << endl;

	system("pause");

	return 0;
}

//------------------------------------
// area(int), function definition

        float area(float dummy)

		{
			float Area;

			Area = 3.1416*(dummy*dummy);

			return Area;

		}
//--------------------------------------

CODE 2

Code:

// average_weight.cpp
// read 10 or n weights and then find average

#include <iostream>
#include <cstdlib>

using namespace std;

int main()
{

        int n;
        float w[n];
        float total_weight=0;

        cout << "how many weights: "; cin >> n;

        for (int i=0; i<n; i++)
        {
                cout << "enter weight #" << (i+1) << ": "; cin >> w[i];
        }

        cout << endl;

        for (int i=0; i<n; i++)
        {
                cout << "entered weight #" << (i+1) << " is: " << w[i] << endl;
        }

        cout << endl;


        for (int i=0; i<n; i++)
        {
                total_weight += w[i];
        }

        float average_weight = total_weight/n;

        cout << "average weight is: " << average_weight << endl;

        system("pause");
        return 0;

}

Problem details given by the crashed window:

Code:

Problem signature:
  Problem Event Name:	APPCRASH
  Application Name:	average_weight.exe
  Application Version:	0.0.0.0
  Application Timestamp:	4de7ebf8
  Fault Module Name:	average_weight.exe
  Fault Module Version:	0.0.0.0
  Fault Module Timestamp:	4de7ebf8
  Exception Code:	c00000fd
  Exception Offset:	000140ce
  OS Version:	6.0.6001.2.1.0.768.11
  Locale ID:	1033
  Additional Information 1:	6199
  Additional Information 2:	bcfd5cce14b739ebe66a3b520492755c
  Additional Information 3:	973d
  Additional Information 4:	a6b61f04fc8b13590bb3150313d0cc3d

Read our privacy statement:
  http://go.microsoft.com/fwlink/?linkid=50163&clcid=0x0409

Link:
http://img855.imageshack.us/img855/4...shedwindow.jpg

Code:

int n;
float w[n];

No no, a thousand times no. You cannot use a variable as the length of a variable-sized array if you don't know what the value of that variable is. Well, you can, as you see, but it is a Very Very Very Bad Thing.

Actually, as you know, it's not standard C++.
So there are two problems:
- It's not standard C++. It's a GNU extension. You should not use it.
- n's value at this point is undefined, and seeing as C++ is a "serial" language, all instructions are executed in turn, meaning that whatever value you hope to put into n later will not affect the creation of the array which happens at an earlier point in time.
What you need to do is change
float w[n] to
std::vector<float> w;
And instead of w[i] = something, do w.push_back(something).

First and foremost use meaningful variable names in your programs, not a,w, etc...

Secondly, it's bad practice to be calling system functions like that and is potentially a security risk. Search the board, i know Prelude explained it pretty well one time. Saying that, here is what's going on:

what you want: What you're trying to do is create a variable length array and the way you've gone about that won't work.

what's happening: Basically when your compiler parses your code and is going to compile it. it needs to know how big to make the w array. Since n has no value, as Elysia says it will result in undefined behaviour, most likely assigning garbege to the variable n. in which case then your trying to use memory you haven't allocated. this is why your program crashes.

what you want to do: is get the value of n from the end user - at your own risk - and then make the array that size. so something like:

Code:

int n;
      
        float total_weight=0;

        cout << "how many weights: "; cin >> n;
        float w[n];

caveat: I'm a C programmer and C99 allows variable length arrays so I'm assuming it should be allowed in C++ since it is a superset of C

here is how it ran on my system using DevC++:

Code:

how many weights: 5
enter weight #1: 1
enter weight #2: 2
enter weight #3: 3
enter weight #4: 4
enter weight #5: 5

entered weight #1 is: 1
entered weight #2 is: 2
entered weight #3 is: 3
entered weight #4 is: 4
entered weight #5 is: 5

average weight is: 3
Press any key to continue . . .

your other option - and the accepted way - is to make n a constant prior to calling main() in your code so:

Code:

define n 45

.....
int main(void){
...

I'm pretty sure the answer is "If we're doing C++, then just use std::vector already."

Originally Posted by tabstop

I'm pretty sure the answer is "If we're doing C++, then just use std::vector already."

lol, I could be wrong but from the OP code, I doubt s/he is there yet. One core dump at a time

Thank you very much, tabstop, Elysia, caroundw5h. You guys are so nice.

Best wishes
Jackson

Originally Posted by caroundw5h

what you want to do: is get the value of n from the end user - at your own risk - and then make the array that size. so something like:

Code:

int n;
      
        float total_weight=0;

        cout << "how many weights: "; cin >> n;
        float w[n];

Like I explained already, this is not standard C++. This is a GNU extension. It is only standard in C99.
The correct thing to do is use std::vector.