Notebook

 Many times I have come across code something like the below: class AClass { public:     std::string getName(); }; std::string AClass::getName() {     return "I am returning"; } Here the member function converts the string constant to std::string while returning. This is in my opinion not necessary. This cast away some flexibility. I mean this way of converting to a std::string while returning to function does not make much sense.  Reason:  If we need in any place to get the name as const char* I need to reverse by calling string::c_str() const char* pName = instanceOfClassA.getName().c_str(); However, keeping the return type as const char* has some benefits: const char* AClass::getName() const {     return "I am returning"; } 1. const char* pName =  instanceOfClassA.getName().getName(); // No Conversion will take place 2. std::string sName = instanceOfClassA.getName().getName(); // Conversion happens to std::string  So, only based ...

The following small program accepts streams from the console and prints out as soon as we hit enter. #include template < typename In, typename Out > void Echo(In& in, Out& out) {     out << in.rdbuf(); } int main() {     Echo(std::cin, std::cout);     return 0; } So, the rdbuf() returns a pointer to the underlying  basic_streambuf  for the input stream and prints it out again on the console. The program runs in a loop until we break it by Ctrl+C.

This is a variation of selection sort, it bring the largest pancake not yet sorted to the top with one flip; take it down to its final position with one more flip; and repeat this process for the remaining pancakes. The algorithm has been discussed in following wiki link: Pancake Sorting The algorithm is visually represented in following link: Pancake Sorting(Visualization) The idea of this sorting is similar to Selection Sort, here taking maximum items one by one and started placing those from last and at the same time reducing the vector size one by one. To place bigger items in order from last using flips two move to their respective positions. Here is the implementation using std::rotate() and std::reverse(). The output: 

This is an effort to implement insertion sort in C++ way using Standard Template Library. The below implementation is for vector of any type. Two STL libraries algorithm has been used. 1. std::lower_bound 2. std::rotate The implementation cross compiler compatible, tested on GNU C++ (5.4.0) and Microsoft C++ (2010). //g++  5.4.0 #include < iostream > #include < algorithm > #include < vector > template< typename T > void insertionSort(std::vector< T > &vec) { typedef typename std::vector< T >::value_type vt; typename std::vector< vt >::iterator itBegin = vec.begin(); typename std::vector< vt >::iterator itEnd = vec.end(); for(typename std::vector< vt >::iterator it = itBegin; it != itEnd; ++it) { typename std::vector ::iterator ins = std::lower_bound(itBegin, it, *it); std::rotate(ins, it, std::next(it)); } } int main() {     int arr[] = {34, 20, 99, 10, 23}; std::vector...

In my last article have seen how we can leverage C++ STL to do selection sort on the vector of integers only. However, the selection sort wasn't very generic to accept vectors of any basic type and sort accordingly. The below program was an effort to make the same selection sort (ascending order) implementation for vectors of any type like vectors of integers, doubles, chars, etc. #include < iostream > #include < algorithm > #include < vector > template < class ForwardIterator1, class ForwardIterator2 > void iters_swap (ForwardIterator1 a, ForwardIterator2 b) {     std::swap (*a, *b); } template < typename T > void selectionSort(std::vector< T > &v) { // Selection sort typedef typename std::vector ::value_type vt; typename std::vector ::iterator it = v.begin(); while(it != v.end()) {     typename std::vector ::iterator i_Min = std::min_element (it, v.end()); iters_swap(i_Min, it); ++it; } } ...

Selection Sort using Standard Template Libraries (not C++ 11): Selection sort works by finding smallest element from the list unsorted list, swap with leftmost element and proceed until all the elements are in order. Selecting the lowest element requires scanning all n elements (this takes n − 1 comparisons) and then swapping it into the first position. Finding the next lowest element requires scanning the remaining n − 1 elements and so on. Hence the complexity is  O(N^2) . The C++ implementation is like below: #include < iostream > #include < algorithm > #include < vector > template < class ForwardIterator1, class ForwardIterator2 > void iters_swap (ForwardIterator1 a, ForwardIterator2 b) {      std::swap (*a, *b); } template < typename T > void selectionSort(T &v) { // Selection sort std::vector ::iterator it = v.begin(); while(it != v.end()) {         std::vec...

A scenario, as per C++ specification supported and also supported by the GNU C++ compiler but not right to do so. Example: The following piece of code works both in the GNU C++ compiler as well as Microsoft C++ compiler. Code execution gives the right result (I mean expected output). #include  < vector > #include < iostream > class Simple { public: Simple(unsigned int aInt) : memberInteger(aInt) { } void Amaze() { myVector = { Simple(1), Simple(2) }; } unsigned int memberInteger; std::vector < Simple > myVector; }; int main() { Simple s(0); s.Amaze(); size_t sz = s.myVector.size(); return 0; } The size of my vector returns 2. Now, I've changed the a code little bit and it gets compiled well in Gnu C++ compiler(4.9.2 and used "g++ -std=c++11 -Wall Simple_Defeat.cpp -o Simple_Defeat" to compile) but not in Microsoft C++ compiler (Checked with 2013): #include < vector > #include < iostream ...