A recursive_depth Function Implementation with Target Type Parameter in C++

Question

This is a follow-up question for A recursive_depth function for calculating depth of nested types implementation in C++. I am rethinking about the proposed recursive_depth function. When it comes to the type std::string, the output is always 1 because the base type is char. However, why it is 1, not 0 from the perspective of encapsulation? Therefore, I am trying to implement another version recursive_depth function with target type in this post. Additional input here is the target base type, the output 0 will be given for the type std::string when target base type is equal to std::string.

The experimental implementation

• recursive_depth template function implementation

  // recursive_depth function implementation with target type
  template<typename T_Base, typename T>
  constexpr std::size_t recursive_depth()
  {
   return std::size_t{0};
  }
  template<typename T_Base, std::ranges::input_range Range>
  requires (!std::same_as<Range, T_Base>)
  constexpr std::size_t recursive_depth()
  {
   return recursive_depth<T_Base, std::ranges::range_value_t<Range>>() + std::size_t{1};
  }
  

Full Testing Code

The full testing code:

// A recursive_depth Function Implementation with Target Type Parameter in C++
#include <algorithm>
#include <array>
#include <concepts>
#include <deque>
#include <iostream>
#include <iterator>
#include <list>
#include <string>
#include <type_traits>
#include <utility>
#include <vector>
#include <ranges>
// recursive_depth function implementation with target type
template<typename T_Base, typename T>
constexpr std::size_t recursive_depth()
{
 return std::size_t{0};
}
template<typename T_Base, std::ranges::input_range Range>
requires (!std::same_as<Range, T_Base>)
constexpr std::size_t recursive_depth()
{
 return recursive_depth<T_Base, std::ranges::range_value_t<Range>>() + std::size_t{1};
}
void recursive_depth_test();
int main()
{
 recursive_depth_test();
}
void recursive_depth_test()
{
 // std::string cases
 std::string test_string = "123";
 std::cout << "Giving base type std::string:" << recursive_depth<decltype(test_string), decltype(test_string)>() << '\n';
 std::cout << "Giving base type char:" << recursive_depth<char, decltype(test_string)>() << '\n';
 
 // non-nested type `char` case
 char test_char = 'A';
 std::cout << "non-nested type `char` case: " << recursive_depth<char, decltype(test_char)>() << '\n';
 
 // non-nested type `int` case
 int test_int = 100;
 std::cout << "non-nested type `int` case: " << recursive_depth<int, decltype(test_int)>() << '\n';
 
 // std::vector<int> case
 std::vector<int> test_vector{ 5, 7, 4, 2, 8, 6, 1, 9, 0, 3 };
 std::cout << "std::vector<int> case: " << recursive_depth<int, decltype(test_vector)>() << '\n';
 // std::vector<std::vector<int>> case
 std::vector<decltype(test_vector)> test_vector2{ test_vector , test_vector , test_vector };
 std::cout << "std::vector<std::vector<int>> case: " << recursive_depth<int, decltype(test_vector2)>() << '\n';
 // std::deque<int> case
 std::deque<int> test_deque;
 test_deque.push_back(1);
 test_deque.push_back(2);
 test_deque.push_back(3);
 test_deque.push_back(4);
 test_deque.push_back(5);
 test_deque.push_back(6);
 std::cout << "std::deque<int> case: " << recursive_depth<int, decltype(test_deque)>() << '\n';
 // std::deque<std::deque<int>> case
 std::deque<decltype(test_deque)> test_deque2;
 test_deque2.push_back(test_deque);
 test_deque2.push_back(test_deque);
 test_deque2.push_back(test_deque);
 std::cout << "std::deque<std::deque<int>> case: " << recursive_depth<int, decltype(test_deque2)>() << '\n';
 // std::list<int> case
 std::list<int> test_list = { 1, 2, 3, 4, 5, 6 };
 std::cout << "std::list<int> case: " << recursive_depth<int, decltype(test_list)>() << '\n';
 // std::list<std::list<int>> case
 std::list<std::list<int>> test_list2 = { test_list, test_list, test_list, test_list };
 std::cout << "std::list<std::list<int>> case: " << recursive_depth<int, decltype(test_list2)>() << '\n';
 std::cout << "std::list<std::list<int>> with given base type std::list<int>: " << recursive_depth<std::list<int>, decltype(test_list2)>() << '\n';
}

The output of the test code above:

Giving base type std::string:0
Giving base type char:1
non-nested type `char` case: 0
non-nested type `int` case: 0
std::vector<int> case: 1
std::vector<std::vector<int>> case: 2
std::deque<int> case: 1
std::deque<std::deque<int>> case: 2
std::list<int> case: 1
std::list<std::list<int>> case: 2
std::list<std::list<int>> with given base type std::list<int>: 1

Godbolt link is here.

All suggestions are welcome.

The summary information:

• Which question it is a follow-up to?

  A recursive_depth function for calculating depth of nested types implementation in C++.

• What changes has been made in the code since last question?

  I am trying to implement another version recursive_depth function with target type in this post.

• Why a new review is being asked for?

  Should it keeps the name recursive_depth or it's better to rename it by something like unwrap_times_to_target_type? Also, other suggestions are welcome.

• chrysante: Why do you construct objects in your test cases if you are only testing static properties of the types?
• Lozminda: get_nested_type_depth() , get_depth() . Possibly type_depth()

Answer 1

What happens in the case the value_type being searched for isn't in the nested set of value_types? Consider recursive_depth<int, std::string>. The value returned will be 2 ([std::basic_string<char>, char]) because int is never encountered. You could also scan the elements of tuple, pair, and ranges::subrange elements. The other 2 tuple-like types are std::array (already handled) and std::complex (undefined behavior if not arithmetic types, unspecified behavior if not standard floating point).

Your tests have run-time operations when you are working on compile-time types. Creating variables and filling them with dummy values doesn't seem within scope.

While it's nice you created readable tests, the cognitive load on the tester to evaluate every test as they read each one is quite burdensome. Calculate the result manually, verify it is correct, program your test with that expected value checked against the observed value. Testing frameworks usually have some sort of static test function that allows you to switch between compile time reporting (hard error) and runtime reporting (onscreen fail message) at the flip of a flag (e.g. Catch2 v3 has STATIC_CHECK).