Reading strings safely

Why reading text is dangerous in C

A program cannot know in advance how long a user's input will be. If it reads into a 20-byte array and the user types 100 characters, a careless function writes the extra 80 bytes over whatever comes next in memory. Attackers deliberately send long inputs to exploit this. Safe input means always telling the reading function how big the buffer is.

fgets: the standard safe way

#include <stdio.h>
#include <string.h>

int main(void) {
    char name[30];
    printf("Full name: ");
    if (fgets(name, sizeof name, stdin) == NULL) {   /* NULL: end of input or error */
        return 1;
    }
    name[strcspn(name, "\n")] = '\0';                 /* remove the newline if present */
    printf("Hello, %s!\n", name);                     /* spaces are kept: "Ravi Kumar" */
    return 0;
}

fgets reads up to size - 1 characters, stops after a newline, and always adds \0. If the line is longer than the buffer, the rest stays in the input for the next read; check whether the newline was found if that matters.

scanf with strings

char city[20];
scanf("%19s", city);          /* reads one word, at most 19 chars; stops at a space */
char line[100];
scanf(" %99[^\n]", line);     /* reads up to a newline, skipping leading whitespace (a "scanset") */

Converting text to numbers: strtol

#include <errno.h>
#include <stdlib.h>

int parse_int(const char *text, long *out) {
    char *end;
    errno = 0;
    long v = strtol(text, &end, 10);         /* base 10 */
    if (end == text) return 0;               /* no digits at all */
    while (*end == ' ' || *end == '\n') end++;
    if (*end != '\0') return 0;              /* junk after the number, e.g. "12kg" */
    if (errno == ERANGE) return 0;           /* too big for long */
    *out = v;
    return 1;
}

atoi("12kg") returns 12 and atoi("abc") returns 0, with no way to know the input was bad, and its behaviour on overflow is undefined. strtol reports where parsing stopped (end) and sets errno to ERANGE on overflow, so you can reject bad input. Use strtod for decimals.

A robust input pattern

1. Read a whole line with fgets into a big enough buffer.
2. Remove the newline.
3. Parse it with strtol / strtod / sscanf and check every result.
4. If it fails, print a clear message and ask again (in a loop) or exit with an error code.