Question

I am relatively new to C++. Recent assignments have required that I convert a multitude of char buffers (from structures/sockets, etc.) to strings. I have been using variations on the following but they seem awkward. Is there a better way to do this kind of thing?

#include <iostream>
#include <string>

using std::string;
using std::cout;
using std::endl;


char* bufferToCString(char *buff, int buffSize, char *str)
{
    memset(str, '\0', buffSize + 1);
    return(strncpy(str, buff, buffSize));
}


string& bufferToString(char* buffer, int bufflen, string& str)
{
    char temp[bufflen];

    memset(temp, '\0', bufflen + 1);
    strncpy(temp, buffer, bufflen);

    return(str.assign(temp));
}



int main(int argc, char *argv[])
{
   char buff[4] = {'a', 'b', 'c', 'd'};

   char str[5];
   string str2;

   cout << bufferToCString(buff, sizeof(buff), str) << endl;

   cout << bufferToString(buff, sizeof(buff), str2) << endl;

}

Answer 1

Given your input strings are not null terminated, you shouldn’t use str… functions. You also can’t use the popularly used std::string constructors. However, you can use this constructor:

std::string str(buffer, buflen): it takes a char* and a length. (actually const char* and length)

I would avoid the C string version. This would give:

std::string bufferToString(char* buffer, int bufflen)
{
    std::string ret(buffer, bufflen);

    return ret;
}

If you really must use the C-string version, either drop a 0 at the bufflen position (if you can) or create a buffer of bufflen+1, then memcpy the buffer into it, and drop a 0 at the end (bufflen position).