weird segmentation fault stemming from couts

Question

so i have this bit of code.. it spews segmentation faults seemingly without rhyme or reason. im doing it for HackerRank practice problems, and when i leave all the couts, it works fine, but i fail because all the extra output.

when i comment out/outright delete the cout lines, i get segmentation faults.

ive written the code in a script form, and a main fxn calling a seperate fxn to output the correct number, and still the same deal. but i know the logic is sound, because it outputs the correct answer at the bottom.

the test is that its inputting 3 lines: 5 on the first, 10 40 30 50 20 on the second, and 1 2 3 4 5 on the third. first line specifies the length of the next 2 arrays, which get filled with the values from the second and third lines. the output is supposed to be only 1 line containing the value 32.0 but i seem to only be able to get it to work with comments that change the output to at least 21 lines

#include <cmath>
#include <cstdio>
#include <vector>
#include <iostream>
#include <algorithm>
using namespace std;
int n;
double wm=0;
vector<double>a1(n);
vector<double>a2(n);

double result(vector<double> arr1, vector<double> arr2){
    double wv=0.0;
    double wv2=0.0;
    for(int i=0; i<n; i++){
        wv+=(arr1[i] * arr2[i]);
        cout<<wv<<"\n";
    }
    for(int i=0; i<n; i++){
        wv2+=arr2[i];
        cout<<wv2<<"\n";
    }
    return wv/wv2;

}
int main() {
    /* Enter your code here. Read input from STDIN. Print output to STDOUT */  
    cin>>n;

    for(int i=0; i<n; i++){
        cin>>a1[i];
        cout<<a1[i]<<"\n";
    } 
    for(int ii=0; ii<n; ii++){
        cin>>a2[ii];
        cout<<a1[i]<<"\n";
    }
    wm=result(a1,a2);
    cout.precision(1);
    cout<<fixed<<wm;
    return 0;
}

*******************THAT IS Main() CALLING Result(), THIS IS THE SCRIPT FORMAT**********

#include <cmath>
#include <cstdio>
#include <vector>
#include <iostream>
#include <algorithm>
using namespace std;

int main() {
    /* Enter your code here. Read input from STDIN. Print output to STDOUT */  
    int n;
    vector<double>a1(n);
    vector<double>a2(n);
    double wv=0.0;
    double wv2=0.0;
    double wm=0;
    cin>>n;

    for(int i=0; i<n; i++){
        cin>>a1[i];
        cout<<a1[i]<<"\n";
    } 
    for(int ii=0; ii<n; ii++){
        cin>>a2[ii];
        cout<<a2[ii]<<"\n";
    }
    for(int i=0; i<n; i++){
        wv+=a1[i] * a2[i];
        cout<<wv<<"\n";
    }
    for(int i=0; i<n; i++){
        wv2+=a2[i];
        cout<<wv2<<"\n";
    }
    wm=wv/wv2;
    cout.precision(1);
    cout<<fixed<<wm;
    return 0;
}

Answer 1

In both codes you are initializing a1 and a2 to size n, which is at that point zero in the first code example and indeterminate (because uninitialized) in the second example. You never resize these vectors, but try to index them later as if they had a larger size after n is read from cin.

This is undefined behavior and whether it will "work" or not is pure luck. In the second example already the initialization with an indeterminate size is undefined behavior as well.

Also beware that initialization order for global variables is tricky, and it only works correctly in this case, because the vectors are initialized in the dynamic initialization phase and n is initialized in the static initialization phase, which is always ordered before the dynamic one. In general there is often no guarantee that global initialization will happen in the expected order.

Some compiler will warn you about the use of an uninitiliazed variable in the second code, if you enable extra warnings and possibly optimizations. (e.g. for clang -Wall -Wextra, for gcc -Wall -Wextra -O2)

Furthermore using namespace std; is bad practice, see this question.