FCRL2 - small factorials

[romgom_sof]

Hi all !!
This is the first good program I try to submit, it works perfectly on my computer, but I always have a "wrong answer" message, please tell me why it is not accepted....
(should I use "fprintf" & fscanf" so that I can read/write the input/output in an external file ???)

Thanks !

#include <stdio.h>
int main(void)
{
    int n,i;
    scanf("%d",&n);
    int *list;
    list=(int*)malloc(n*sizeof(int));
    double a;
    for(i=0;i<n;i++)  scanf("%d",&list[i]);
    for(i=0;i<n;i++)
    {
        if(list[i]==0)
        printf("1\n");
        else
        {
            a=list[i];
            while(list[i]>1)
            {
                a=a*(list[i]-1),
                list[i]--;
            }
            printf("%1.f\n",a);
        }
    }
    return 0;
}

[TripleM_sof]

Just because a double can store bigger numbers than an integer, that doesn't mean it is more accurate.

If you test your program on the largest input, you'll see that it obviously gives the wrong answer.

[romgom_sof]

Yes, but it's said that we have to calculate n! for 1<=n<=100, and it works well for all this cases ....

[TripleM_sof]

Uh, no it doesn't. As I said, test it for n=100, and you'll see it obviously doesn't work.

[romgom_sof]

I'm sorry, but I tried, and even for 100!, I have a good result (I checked it, I wouldn't have posted otherwise)
I obtain 933262154439441750000000....000 (or 9,33.. * 10^157)

So how shall I do if you think I shouldn't use "double" type ?

[TripleM_sof]

Those 0s should have given it away - sure, it may end with about 25 zeroes, but not 150 of them!

You'll need to use some sort of big integer class. (arbitrary-precision arithmetic).

[romgom_sof]

Thx, I understand now where is my problem

[kokosek_sof1]

Zrób może zmienną globalną tab2 i zamiast putchara daj: tab2[i]=tab[i];