String parsing in C. Copying a part of a string to another string

Question

I'm trying to write a function which would convert all escape sequences in a string in their non-printable form. Basically if I have a string "This \n makes a new line", I would like it to be "This makes a new line". So far I've got this. I'm calling from main:

int main()
{
    unescape("This \\n\\n is \\t\\t\\t string number \\t 7.");
    return 0;
}

char* unescape(char* s)
{
    char *esc[2] = {"\\n", "\\t"};
    int i;
    char* uus = (char*)calloc(80, sizeof(char));
    char* uus2 = (char*)calloc(80,sizeof(char));

    strncpy(uus, s, strlen(s));

    for(i = 0; i < 2; i++)
    {
        while(strstr(uus, esc[i]) != NULL) //checks if \\n can be found
        {
            //printf("\n\n%p\n\n", strstr(uus, esc[i]));
            int c = strstr(uus, esc[i]) - uus; //gets the difference between the address of the beginning of the string and the location
                                           //where the searchable string was found
            uus2 = strncpy(uus2, uus, c); //copies the beginning of the string to a new string

            //add check which esc is being used
            strcat(uus2, "\n"); //adds the non-printable form of the escape sequence
            printf("%s", uus2);

            //should clear the string uus before writing uus2 to it 
            strncpy(uus, uus2, strlen(uus2)); //copies the string uus2 to uus so it can be checked again
        }
    }
    //this should return something in the end. 
}

Basically, what I need to do now, is take the part from the string uus after "\n" and add it to the string uus2 so I can run the while loop again. I thought about using strtok but hit a wall as it makes two separate strings using some kind of delimiter which is not always there in my case.

edit: Adding the rest of the string to uus2 should be before strncpy. This is the code without it.

edit vol2: This is the code that works and which I ended up using. Basically edited Ruud's version a bit as the function I had to use had to return a string. Thanks a lot.

char* unescape(char* s)
{
    char *uus = (char*) calloc(80, sizeof(char));
    int i = 0;

    while (*s != '\0')
    {
        char c = *s++;
            if (c == '\\' && *s != '\0')
            {
                c = *s++;
                switch (c)
                {
                case 'n': c = '\n'; break;
                case 't': c = '\t'; break;
                }
            }
        uus[i] = c;
        i++;
    }
    uus[i] = '\0';
    return uus;
}

There were at least 3 problems with your original code: (1) you were so busy copying data back and forth that you forgot to close the 'holes' that these escape characters leave behind; (2) you left strings unterminated because strncpy never copied the trailing \0; (3) \t was replaced by a newline because of the hardcoded \n in strcat(uus2, "\n");. — Ruud Helderman
– Ruud Helderman, Commented Apr 12, 2014 at 21:16

Ruud Helderman · Accepted Answer · 2014-04-12 20:01:07Z

2

I agree with Anonymouse. It is both clumsy and inefficient to replace first all \n, then all \t. Instead, make a single pass through the string, replacing all escape characters as you go.

I left the space allocation out in the code sample below; IMHO this is a separate responsibility, not a part of the algorithm, and as such does not belong in the same function.

void unescape(char *target, const char *source)
{
    while (*source != '\0')
    {
        char c = *source++;
        if (c == '\\' && *source != '\0')
        {
            c = *source++;
            switch (c)
            {
                case 'n': c = '\n'; break;
                case 't': c = '\t'; break;
            }
        }
        *target++ = c;
    }
    *target = '\0';
}

EDIT: Here's an alternative version, using strchr as suggested by Anonymouse. This implementation should be faster, especially on very long strings with relatively few escape characters. I posted it primarily as a demonstration of how optimizations can make your code more complex and less readable; and consequently less maintainable and more error-prone. For a detailed discussion, see: http://c2.com/cgi/wiki?OptimizeLater

void unescape(char *target, const char *source)
{
    while (*source != '\0')
    {
        if (*source++ == '\\' && *source != '\0')
        {
            char c = *source++;
            switch (c)
            {
                case 'n': c = '\n'; break;
                case 't': c = '\t'; break;
            }
            *target++ = c;
        }
        else
        {
            const char *escape = strchr(source--, '\\');
            int numberOfChars = escape != NULL ? escape - source : strlen(source);
            strncpy(target, source, numberOfChars);
            target += numberOfChars;
            source += numberOfChars;
        }
    }
    *target = '\0';
}

edited Apr 12, 2014 at 20:01

answered Apr 12, 2014 at 16:43

Ruud Helderman

11.1k1 gold badge28 silver badges49 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

RubberDuckDebugging Over a year ago

Going to try this out. Though, just being curious, is there a way to pull my approach off? Just wondering for future reference as I've spent so much time trying to make it happen that way.

Anonymouse Over a year ago

my version with strchr will be faster. Plus I'm (mistakenly) searching \\n - so delete the IF statement. Finally Ruud's function prototype is better than the questioners, const is your friend especially if you try and modify a constant string (such as "This \\n\\n is \\t\\t\\t string number \\t 7.")

RubberDuckDebugging Over a year ago

I ended up using Ruud's version of the code. I edited my first post with the final code. Sadly, I had some trouble understanding Anonymouse's code but thanks a lot to both of you.

Ruud Helderman Over a year ago

@Anonymouse: Indeed, strchr is faster, but speed isn't everything. Please see my edit; and the wiki discussion I hyperlinked.

Anonymouse · Accepted Answer · 2014-04-12 16:37:11Z

1

You'd be better using this...

char *p;

p = input_string;
while ((p=strchr (p, '\\')) != NULL)
{
  if (p [1] == '\\')
  {
     switch (p [2])
     {
     case 'n' :
       // handle \n
       break;
     case 't' :
       // handle tab
       break;    
     }
  }
  else
    p++;

}

answered Apr 12, 2014 at 16:37

Anonymouse

95510 silver badges22 bronze badges

Collectives™ on Stack Overflow

String parsing in C. Copying a part of a string to another string

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related