Thread: C++ string contents

Code:

#include <iostream>
#include <string>
using namespace std;

int main() {
    string str = "Hello\0World";
    const char *c_str = str.c_str();

    for(int i = 0; i < 20; i++) {
        cout << i << ": " << *c_str++ << endl;
    }

    return 0;
}

I've noticed with this program that a C++ string won't include anything after the null character when creating a C string. But does the C++ string include the "world" substring to itself? Does the C++ string even include the '\0' to itself?

I believe the answer to your question is "it depends". The internal storage of a standard-class is not defined beyond the obvious - if it says that it's storing a string, it will store a string - but there's nothing saying how the string is actually stored - for all we know, "Helllo, World" could be stored as a linked list of one character per node, or it could be stored backwards. c_str() may well return a pointer to a different piece of data than the "actual string".

Of course, most of what I describe is pretty unrealistic - but the fact is that YOU CAN'T RELY on what's inside a string class unless you write it yourself (besides the defined functionality of the functions and operators defined by the external class declaration of course). It may vary depending on which compiler/library you use. It can change from one version of compiler/library to another.

It's quite possible that the string length is stored, and _UNLESS_ you call c_str(), there is no zero at the end of the string. Note that the c_str() returns a const - you can't CHANGE the content of the string with the returned value, so it's entirely possible for the implementation to make a duplicate of the "actual string" and return the address of this duplicate as the result to c_str()....

--
Mats

I think I sort of misread you question: There's a NUL character in the middle of your string, between Hello and World - this terminates the C-string when used to construct the C++-string. It's not ever seeing the "world" bit of that string (it's not that it couldn't, if the constructor wanted to read beyond the '\0' - but what would happen if you actually just intended it to contain "Hello\0" and that is the very last byte of a valid memory region, with the very next address in memory resulting an invalid memory access?)

So, the conclusion is: You can't store zero's inside a C++-string when you construct it with a string - it is possible that you can store a zero inside the string at other points (e.g. using the array operator method) - but it's actually fairly unlikely that this would be useful.

--
Mats

Originally Posted by CornedBee

You can store them when you construct with a string, but you have to pass the length explicitly.

Ah, ok, (and thanks for the clarficiation on the data() - whcih in the implementation I looked at actually calls "c_str()"). So one could do

Code:

string("Hello\0world", 11);

--
Mats

so how would you get the full string back into a C-string?

Originally Posted by robwhit

so how would you get the full string back into a C-string?

You can't, if it contains zero bytes. By definition the zero byte terminates the C-string.

well ok, an array.

Originally Posted by robwhit

well ok, an array.

The data() method returns a pointer to the data. Along with the size returned by size(), you can use this to get at the data.

thanks.

And if you want your own copy, then use memcpy to copy it, not strcpy which would stop at the first embedded null. You might be able to use strncpy, I'm not sure, but I'd stick with memcpy because you aren't copying a C string, so you shouldn't be using C string functions.

Remember that if you want to modify the data, you need your own copy. You're not really allowed to modify the return value of data().

Thanks for the replies.