Bubble Sort Implementation | C-ish and Modern C++

Question

For practice purpose, I wrote bubble sort in C++ two styles: C-ish and modern C++ compliant. I would like to have your comments on any point on these implementations.

Sorting.h

#ifndef SORTING_ALGORITHMS_H
#define SORTING_ALGORITHMS_H
#include <utility>
#include <vector>
//#include <iterator>
#include <algorithm>
namespace etpc
{
 template <class T>
 void sortBubble(T* pArrHead, std::size_t sArrSize)
 {
 for(std::size_t i=0; i<sArrSize-1; i++)
 {
 for(std::size_t j=0; j<sArrSize-i-1; j++)
 {
 if(pArrHead[j]>pArrHead[j+1])
 std::swap(pArrHead[j], pArrHead[j+1]);
 }
 }
 }
 // May 2, 2020
 // Based on the following link
 // https://rosettacode.org/wiki/Sorting_algorithms/Bubble_sort#C.2B.2B
 // typename vs class
 // https://stackoverflow.com/questions/2023977/difference-of-keywords-typename-and-class-in-templates
 template <class T>
 void sortBubble(std::vector<T>& vArrHead)
 {
 typename std::vector<T>::iterator begin = std::begin(vArrHead);
 typename std::vector<T>::iterator end = std::end(vArrHead);
 while (begin != end--) {
 for (auto it = begin; it != end; ++it) {
 if (*(it + 1) < *it) {
 std::iter_swap(it, it + 1);
 }
 }
 }
 }
}
#endif // SORTING_ALGORITHMS_H

Main.cpp

#include "Sorting/include/Sorting.h"
#include <iostream>
#include <chrono>
#include <iterator>
int main()
{
 std::chrono::steady_clock::time_point begin;
 std::chrono::steady_clock::time_point end;
 std::ostream_iterator<int> out_it (std::cout,", ");
 static const int i32Size = 50;
 int arr[i32Size] = {7804, 50398, 14945, 1814, 51383, 63156, 8432, 58103, 28175, 4339, 8361, 37158, 1529, 43066, 62052, 9591, 13168, 332, 55913, 2418, 48066, 46504, 52922, 39523, 36653, 30402, 9373, 56202, 50539, 41187, 42606, 32278, 63902, 41668, 7505, 46534, 25846, 49739, 63411, 45933, 15042, 6544, 35718, 17035, 34647, 15212, 52690, 64299, 61535, 45071};
 begin = std::chrono::steady_clock::now();
 etpc::sortBubble<int>(arr, i32Size);
 end = std::chrono::steady_clock::now();
 std::cout << "Time difference impArray = " << std::chrono::duration_cast<std::chrono::nanoseconds>(end - begin).count() << "[ns]" << std::endl;
 std::copy(std::begin(arr), std::end(arr), out_it);
 std::cout << '\n';
 std::vector<int> v = {7804, 50398, 14945, 1814, 51383, 63156, 8432, 58103, 28175, 4339, 8361, 37158, 1529, 43066, 62052, 9591, 13168, 332, 55913, 2418, 48066, 46504, 52922, 39523, 36653, 30402, 9373, 56202, 50539, 41187, 42606, 32278, 63902, 41668, 7505, 46534, 25846, 49739, 63411, 45933, 15042, 6544, 35718, 17035, 34647, 15212, 52690, 64299, 61535, 45071};
 begin = std::chrono::steady_clock::now();
 etpc::sortBubble<int>(v);
 end = std::chrono::steady_clock::now();
 std::cout << "Time difference impVector = " << std::chrono::duration_cast<std::chrono::nanoseconds>(end - begin).count() << "[ns]" << std::endl;
 std::copy(std::begin(v), std::end(v), out_it);
 std::cout << '\n';
 return 0;
}
```

• JVApen: You can still replace it +1 by std::next(it)
• JVApen: It's unclear to me what you are trying to get out of this review, can you elaborate?
• Erdem Tuna: For C-style implementation, coding practises, algorithm-specific recommendation or your personal comments. For modern C++ implementation, does it suit with modern C++, can it be improved etc? @JVApen

Answer 1

As you're already working with iterators you might as well accept iterators as arguments instead of the std::vector (as all STL algorithms do). That way it will work with other containers as well, even with raw pointers.

The comparison can also be customized like in std::sort. Maybe the caller wants to sort in descending order or compare structures by some field.