Is there a way to mmap a stdin? - c

Question

I have a following problem:
My job is to write a program that takes unsigned integer numbers that are passed to it through stdin and print out only the numbers that have more than 2 bits set to one. How should I do it efficiently? I did a version of a program where I read the numbers from a file using mmap, and it's quite quick. I read it like a very big *char buffer and using strtol I 'scrub' out each number and do my check and whatnot.
Is there a way to operate on a string passed through stdin the same way? I though about buffering using fread, but there is a problem, where the buffer cuts off the number (meaning if i pass "1024 35" and I have a 6 byte buffer I will get "1024 3"), and I shudder to think how to get around that.
Source:
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/mman.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h> /* mmap() is defined in this header */
#include <fcntl.h>
#include<string.h>
#include"apue.h"
int main (int argc, char *argv[])
{
int fdin, fdout;
char *src, *dst;
struct stat statbuf;
/* open the input file */
if ((fdin = open (argv[1], O_RDONLY)) < 0)
{printf("can't open %s for reading", argv[1]);return 1;}
/* find size of input file */
if (fstat (fdin,&statbuf) < 0)
{printf("fstat error");return 1;}
/* mmap the input file */
if ((src = mmap (0, statbuf.st_size, PROT_READ, MAP_SHARED, fdin, 0))
== (caddr_t) -1)
{printf("mmap error for input");return 1;}
char* beg=src;
long x;
char* end=&src[statbuf.st_size-1];
while(src<end)
{
beg=src;
x = strtol (src,&src,10);
if(!((x != 0) && ((x & (~x + 1)) == x)))
fwrite(beg, 1, (int)(src-beg), stdout);
}
return 0;
}
http://pastebin.com/EVhG3x79

Answer 1

I think the expected solution is how to count the ones and not how to read from stdin.
int count_ones(int n);
means the question is how to implement the count_ones efficiently.
and your main just should look like this:
int main()
{
int x;
cin>>x;
if( count_ones(x)>2){
cout<<x<<endl;
}
return 0;
}
I think the expected answer is:
use array size 256
for each byte(=unsigned char) put in the array in its place the count of ones (can be: from 0 to 8)
split each number to its bytes and sum the ones on each of the bytes.
return the result