Creating my own strcmp () function in C

问题:

I was assigned by my teacher to write my own strcmp() function in C. I did create my own version of said function, and I was hoping to get some feedback.

int CompareTwoStrings ( char *StringOne, char *StringTwo ) {
    // Evaluates if both strings have the same length.
    if  ( strlen ( StringOne ) != strlen ( StringTwo ) ) {
        // Given that the strings have an unequal length, it compares between both
        // lengths.
        if  ( strlen ( StringOne ) < strlen ( StringTwo ) ) {
            return ( StringOneIsLesser );
        }
        if  ( strlen ( StringOne ) > strlen ( StringTwo ) ) {
            return ( StringOneIsGreater );
        }
    }
    int i;
    // Since both strings are equal in length...
    for ( i = 0; i < strlen ( StringOne ); i++ ) {
        // It goes comparing letter per letter.
        if  ( StringOne [ i ] != StringTwo [ i ] ) {
            if  ( StringOne [ i ] < StringTwo [ i ] ) {
                return ( StringOneIsLesser );
            }
            if  ( StringOne [ i ] > StringTwo [ i ] ) {
                return ( StringOneIsGreater );
            }
        }
    }
    // If it ever reaches this part, it means they are equal.
    return ( StringsAreEqual );
}

StringOneIsLesser, StringOneIsGreater, StringsAreEqual are defined as const int with the respective values: -1, +1, 0.

Thing is, I'm not exactly sure if, for example, my StringOne has a lesser length than my StringTwo, that automatically means StringTwo is greater, because I don't know how strcmp() is particularly implemented. I need some of your feedback for that.

回答1:

So much for such a simple task. I believe something simple as this would do:

int my_strcmp(char *a, char *b)
{
    while (*a && *b && *a == *b) { ++a; ++b; }
    return (unsigned char)(*a) - (unsigned char)(*b);
}

回答2:

strcmp compares alphabetically: so "aaa" < "b" even though "b" is shorter.

Because of this, you can skip the length check and just do the letter by letter comparison. If you get to a NULL character while both strings are equal so far, then the shorter one is the lesser one.

Also: make StringsAreEqual == 0, not 1 for compatibility with standard sorting functions.

回答3:

Try this also for your better understanding:

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

int main(void)
{
    char string1[20], string2[20];
    int i=0,len=0, count=0;
    puts("enter the stirng one to compare");
    fgets(string1, sizeof(string1), stdin);
    len = strlen(string1);
    if(string1[len-1]=='\n')
    string1[len-1]='\0';

    puts("enter the stirng two to compare");
    fgets(string2, sizeof(string2), stdin);
    len = strlen(string2);
    if(string2[len-1]=='\n')
    string2[len-1]='\0';
    if(strlen(string1)==strlen(string2))
    {
    for(i=0;string1[i]!='\0', string2[i]!='\0', i<strlen(string1);i++)
    {
        count=string1[i]-string2[i];
        count+=count;
    }
        if(count==0)
            printf("strings are equal");
        else if(count<0)
            printf("string1 is less than string2");
        else if(count>0)
            printf("string2 is less than string1");
    }

    if(strlen(string1)<strlen(string2))
    {
    for(i=0;string1[i]!='\0', i<strlen(string1);i++)
    {
        count=string1[i]-string2[i];
        count+=count;
    }
        if(count==0)
            printf("strings are equal");
        else if(count<0)
            printf("string1 is less than string2");
        else if(count>0)
            printf("string2 is less than string1");
    }

    if(strlen(string1)>strlen(string2))
    {
    for(i=0;string2[i]!='\0', i<strlen(string2);i++)
    {
        count=string1[i]-string2[i];
        count+=count;
    }
        if(count==0)
            printf("strings are equal");
        else if(count<0)
            printf("string1 is less than string2");
        else if(count>0)
            printf("string2 is less than string1");
    }


    return 0;
}

回答4:

    int mystrncmp(const char * str1, const char * str2, unsigned int n)
     {
      while (*str1 == *str2) {
          if (*str1 == '\0' || *str2 == '\0')
             break;

          str1++;
          str2++;
       }


   if (*str1 == '\0' && *str2 == '\0')
      return 0;
   else
      return -1;
}

回答5:

strcmp() is fairly easy to code. The usual mis-codings issues include:

Parameter type

strcmp(s1,s2) uses const char * types, not char *. This allows the function to be called with pointers to const data. It conveys to the user the function's non-altering of data. It can help with optimization.

Sign-less compare

All str...() function perform as if char was unsigned char, even if char is signed. This readily affects the result when strings differ and a character outside the range [1...CHAR_MAX] is found.

Range

On select implementations, the range of unsigned char minus unsigned char is outside the int range. Using 2 compares (a>b) - (a-b) avoids any problem rather than a-b;. Further: many compilers recognized that idiom and emit good code.

int my_strcmp(const char *s1, const char *s2) {
  // All compares done as if `char` was `unsigned char`
  const unsigned char *us1 = (const unsigned char *) s1;
  const unsigned char *us2 = (const unsigned char *) s2;

  // As long as the data is the same and '\0' not found, iterate
  while (*us1 == *us2 && *us1 != '\0') {
    us1++;
    us2++;
  }

  // Use compares to avoid any mathematical overflow 
  // (possible when `unsigned char` and `unsigned` have the same range).
  return (*us1 > *us2) - (*us1 < *us2);
}