Is there any rules in the standard that violates modifications of internal std::string buffer returned by operator[] like this:
void foo(char* buf)
{
buf[1] = 's';
}
std::string str = "str";
modify_buffer(&str[0]);
I found the following quote in the C++11 draft about data and c_str functions:
Requires: The program shall not alter any of the values stored in the character array.
But I don't see any about operator[].
operator[]
operator[] returns a reference to the character. So if the string is NOT const, you can modify it safely.
For C++ 11, the characters are stored contiguously, so you can take &str[0] as the beginning of the underlying array whose size is str.size(). And you can modify any element between [ &str[0], &str[0] + str.size() ), if the string is NOT const. e.g. you can pass &str[0] and str.size() to void func(char *arr, size_t arr_size): func(&str[0], str.size())
data()andc_str()members
For C++11 and C++14, both data() and c_str() returns const CharT*, so you CANNOT modify element with the returned pointer. However, from C++17, data() will return CharT*, if string is NOT const. And data() will be an alias to &str[0].
char elements in buckets, so don't generally expect them all to be consecutive. I have written a String (beware that i don't write string) with same semantics as string but with copy on write specifications, and chars where contiguous up to some point only.
std::stringin your example.data()andc_str()returnsconst char*operator[]then why do you think a pointer to the element would be any different?