Thread: read more chars than the buffer side and detect while preventing overflow

I want to check is the user inputs a word that has more than 10 characters.If he does,let's say we exit the program.

This not my code,so do not comment the global variables etc...

Code:

#include <assert.h>
#include <string.h>
#include <stdio.h>
#include <wchar.h>
#include <locale.h>

typedef struct {
    wchar_t term[10];
    int freq;
} word_t;

/* global array of words */
word_t words[5000];
int n_words = 0;

int find_word(wchar_t s[]) {
    int i;
    for (i=0; i < n_words; i++)
        if (wcsncmp(s, words[i].term,10) == 0)
            return i;
    return -1;
}

void add_word(wchar_t s[]) {
    assert(wcslen(s) <= 10);
    wcsncpy(words[n_words].term, s,10);
    words[n_words].freq = 1;
    n_words++;
}

int main(void) {
    char myStr[11];
    wchar_t s[11];
    setlocale(LC_CTYPE,"UTF-8");
    printf("Please type a word with 10 characters at most\n");
    /*could use fgetws and wprintf instead*/
    while ((fgets(myStr,10,stdin)!= NULL)) {
        if(strlen(myStr)>10)
        {
             printf("Word with more than 10 characters detected, so exiting...\n");
             return -1;
        }
        mbstowcs(s,myStr,strlen(myStr));
        int i = find_word(s);
        if (i == -1)
        {
            add_word(s);
            printf("1: %s\n", myStr);
        }
        else
        {
            words[i].freq++;
            printf("%d: %s\n",words[i].freq ,myStr);
        }
        printf("Please type a word with 10 characters at most\n");
    }
    return 0;
}

When the user enters a word larger than 10 characters,then this word will split at the 10-nth character because of the fgets (with scanf and %10 it will be the same i guess),so it will re-loop without getting anything anything from input ,reading the rest of the word (if it is less than 20 characters).
I thought to go dynamically and ask the user for the size of the buffer but then again i have to make sure that the user will not overflow my little buffer.
Also an idea of declaring a larger array before, breaks my heart because it is such a waste of memory.
So maybe i could look at the end of the loop and determine if there is some letters left in the stdin buffer.If so this means that the user typed a bigger word than 10 characters.
Any ideas?

Depending on how large the array, I wouldn't worry. Unless this is some embedded or otherwise limited-resource system, 500-1k bytes is no biggie (I often use BUFSIZ, it's declared in stdio.h).

Perhaps you could just use getchar() in a loop, check for newline/EOF yourself, and check your 10 char limit.

Originally Posted by anduril462

Depending on how large the array, I wouldn't worry. Unless this is some embedded or otherwise limited-resource system, 500-1k bytes is no biggie (I often use BUFSIZ, it's declared in stdio.h).
Perhaps you could just use getchar() in a loop, check for newline/EOF yourself, and check your 10 char limit.

Exactly what i just thought it thanks

fgets will never overflow a buffer on purpose, you the coder have to pass in bad arguments to make it happen, so the check does not even make sense right now.

I don't actually have a better answer for you. You already know that the word will break off. I'm not seeing the problem with it though. I mean, the same word will break off in the same place. Take any sufficiently long word, like, mentholated-cigarettes. Mentholat ed-cigare ttes ... It would occur as frequently as the entire actual word.