Question

I’ve decided to come back to C++ after some time spent in Java and now I’m quite confused about how strings work in C++.

To start with, suppose we have a function:

void fun() {
   int a = 1;
   Point b(1,2);
   char c[] = "c-string";
}

As I understand, a and b are allocated on the stack. c (the pointer) is allocated on the stack too, but the contents (“c-string”) live happily on the heap.

Q1: Are the contents of c automatically deallocated when the function fun ends?

Secondly let’s suppose we have a c++ string:

void fun2() {
  (1) string s = "c++ string";
  (2) s += "append";
  (3) s = "new contents";
  (4) s = "a" + s + "c";
}

String documentation isn’t too specific about how the strings work, so here are the questions:

Q2: Are the contents of s automatically deallocated after fun2 ends?

Q3: What does happen when we concatenate two strings? Should I care about memory usage? (line 2)

Q4: What happens when we overwrite the contents of a string (line 3) – what about memory, should I worry? Is the originally allocated space reused?

Q5: What if I construct a string like this (line 4). Is it expensive? Are string literals ("a","c") pooled (like in Java) or repeated throughout the final executable?

What I am ultimately trying to learn is how to correctly use strings in C++.

Thanks for reading this,
Queequeg

Answer 1

As I understand, a and b are allocated on the stack. c (the pointer) is allocated on the stack too, but the contents (“c-string”) live happily on the heap.

That’s wrong, they all live in automatic memory (the stack). Even the char array. In C++, a string is an object of type std::string.

Q1: Are the contents of c automatically deallocated when the function fun ends?

Yes.

Q2: Are the contents of s automatically deallocated after fun2 ends?

Yes.

Q3: What does happen when we concatenate two strings? Should I care about memory usage? (line 2)

They are concatenated, and the memory is managed automatically. (assuming we’re talking about std::string and not char[] or char*.

Q4: What happens when we overwrite the contents of a string (line 3) – what about memory, should I worry? Is the originally allocated space reused?

Implementation detail. It can be reused, it can be re-allocated if the previous memory can’t hold the new contents.

Q5: What if I construct a string like this (line 4). Is it expensive? Are string literals (“a”,”c”) pooled (like in Java) or repeated throughout the final executable?

String literals can be pooled, but it’s not required. For large concatenation, it’s usual to use a std::stringstream instead (similar to Java). But profile first, don’t do premature optimizations. Not that neither of those are string literals though.

char* pStr = "this is a string literal";

This resides in read-only memory and can’t be modified.

What I am ultimately trying to learn is how to correctly use strings in C++.

Use a std::string.