Is `c_str()` in C 11 still guaranteed to return a null-terminated string?

c_str() and Null-Terminated Strings in C 11

In C 11, the definition of basic_string::c_str() raises a question: does it still produce a null-terminated string?

The Issue

In C 11, basic_string::c_str is defined to align with basic_string::data, which in turn aligns with (begin() n) and (&*begin() n). There is no explicit requirement for a null character at the end of the string.

The Answer

Despite the lack of explicit definition, strings in C 11 are required to use null-terminated buffers internally. This is evident in the definition of operator[]:

"Returns: *(begin() pos) if pos < size(), otherwise a reference to an object of type T with value charT(); the referenced value shall not be modified."

Since c_str is defined in terms of operator[], and both c_str and data must be O(1), implementations are forced to utilize null-terminated buffers.

Additional Considerations

The return value requirement for operator[] implies that you can use &operator[](0) to retrieve c_str(). This means that the terminating null character must reside in the same buffer. Consequently, even if the terminator is initialized lazily, it cannot be observed in an intermediate state.

Conclusion

Contrary to initial assumptions, c_str() in C 11 is still guaranteed to produce a null-terminated string, ensuring compatibility with legacy code and enabling efficient character access.

The above is the detailed content of Is `c_str()` in C 11 still guaranteed to return a null-terminated string?. For more information, please follow other related articles on the PHP Chinese website!