C中带有malloc的子字符串函数返回不正确的值

Question

I am implementing a parser in C for first order logic formulas. To check for a binary connective formula (ie in the form (A BinaryConnective B)) I need to split the string and check if A and B are formulas. I have done this using a subString function and I call it from partone and parttwo (A and B respectively):

char *partone(char *g) {
    //Given the formula (A*B) this returns A
    return subString(g, 1, binPosition(g));
}

char *parttwo(char *g) {
    //Given the formula (A*B) this returns B
    return subString(g, binPosition(g) + 1, strlen(g) - 1);
}

The substring function is the following:

char *subString(char *g, int start, int end) {
    //the substring includes index start but does not include the end index.
    char *substr = malloc(sizeof(char)*(end - start));
    int i;
    for(i = 0; i < (end - start); i++) {
        substr[i] = g[start + i];
    }
    return substr;
}

This works when I pass any function except a negated formula (we are using the character "-" to denote negation). For example when I pass (-X[xz]>X[yz]), the program returns "Not a Formula", but if I write the same without the negation it works perfectly. The problem is that the substr that is returned for partone() is "-X[xz]$" where $ can be any random character that I think was stored in memory before. Any ideas why this is happening ONLY in this case? I am new to C and I have looked everywhere.

Thanks in advance.

Answer 1

You forgot the terminating NUL in your subString function. Your malloc should look like

char* substring = (char*)malloc (sizeof (char)*(end-start+1));

And at the end you need to terminate the result with 0:

substring [end-start]=0;

And since arrays in c are zero-indexed, I'm suprised that you don't lose the leading "-" of part A, since partone calls subString with 1 as start .

EDIT: there are a lot of built-in functions for string manipulation in C/C++. You should use them since they are tested and optimised. I did not use C so much for many years now so I'm not up to date what are the appropriate methods in C++11 today, but Google may tell you.

EDIT: As chux pointed out, the meaning of your end parameter is not fully clear. So you need to check the strlen (g)-1 in parttwo or the loop conditions in subString and increase to +2 in my malloc statement`.

Answer 2

OP's use of start and end is not formalized, but I'll assume they are indexes into the string of the first and last character of the desired sub-string.

Recall in C, array indexes begin with 0 as in g[0] is the first character of the string stored at g .

If that case, code needs to allocate end - start + 1 + 1 space and insure the allocated character array has a null character '\\0' at the end.

char *subString(const char *g, int start, int end) {
  if (end < start || start < 0) {
    return NULL;
  }
  size_t size = 2u + end - start;
  char *substr = malloc(size);
  if (substr) {
    int i;
    for(i = 0; i < (end - start); i++) {
      substr[i] = g[start + i];
    }
    substr[i] = '\0';
  }
  return substr;
}

Other simplifications/improvements possible