Skip to content

Latest commit

 

History

History
578 lines (505 loc) · 20.3 KB

README.md

File metadata and controls

578 lines (505 loc) · 20.3 KB

pse library

FAQ

What is the pse library?

  • Well, it is a library that does not depend on the standard library! It is still a work in progress with lots of features that still need to be implemented. But it resembles lots of features that the standard library has.

How can I contribute?

  • There are many ways that you can contribute. First, by trying to fix issues that have been put in the issues section of this repository. Secondly, by implementing a feature that you think would be a good feature in the pse library and then opening a pull request!

How old is the pse library?

  • Really new! It started in late November 2022!

Contributors

A list of all the people have helped in the process of making this repository!

The Documentation:

Total Number of Lines: 4609 Last Checked: 12/23/2022 17:18 PM London.

Parse is a helpful open-source library for C++. Consisting of many features that are part of the C++ Standard Library Here are some examples:

A pse::vector is just like a std::vector, consisting of all of the same functionality.

template<typename T>
struct Vector{};

Tuple is again just like the std::tuple with all of the functionality.

template<typename T, typename... Ts>
struct Tuple{};

But this one is currently in progress. I am currently now making one that doesn't increase in size but decreases in size only, so it has part of the functionality of a vector but not the push_back functionality part. It will be called DecreasingTuple

template<typename T, typename... Ts>
struct TupleLike
{
  template<typename T>
  auto push_back(B b);
  
  auto operator[](int i);
}; // Just Tuple combined with a vector

Notice the operator[], it is novel, but bear in mind it only works for tuple of the same types. This is because in order to have varying return types it has to be done at comile time, and passing an arguement to a function like this: foo(int x) is not done at compile time, therefore it will always return the type of the first type in the parameter pack!

swap is also just like std::swap. But a bit more basic.

template<typename T>
auto swap(T& a, T& b)
{
  // Implementation
}

There is also swap for pse::Iterator using pse::Itertor.get_ref and swapping like that. so if you had: pse::Iterator<int> it1{int42pointer}; and pse::Iterator<T> it2{int24pointer}; and then we called pse::swap that this time is implemented in the Algo.cpp header file, and did std::cout << it1.get_val(); and std::cout << it2.get_val(); the output would be 2442. Notice this is because we didn't output \n. It looks the same as pse::swap but the generic version.

// Algo.cpp
template<typename T>
auto swap(Iterator<T>& it)
{
  // swaps the two values
}

Others also include:

pse::Range:

This is just a structure like a Initializer_List in that it stores data as a list with member functions. But it can also hold different types of types like int and float but it is not used for this purpose it is used for being like an Initialiser_List. This Range structure also has member functions, more are still to be implemented but the major one for now is pse::Range::Take(int a, int b) it takes two integers and returns the bit of the array in between those two indexes. There is not any error checking yet but will be soon.

template<typename T, typename... Ts>
struct Range
{
  // Implemenetation.
};

There is also pse::Iterator which looks like the following:

template<typename T>
struct Iterator
{
  auto operator++();
  auto operator--();
  auto operator+=();
  auto operator-=();
  
  auto get();
  auto get_val();
  decltype(auto) get_ref();
  
  auto get_Increment();
  auto get_Dectrement();
  
  auto get_PlusEquals();
  auto get_MinusEquals();
  
  auto begin();
  auto end();
  
  int size;
};

It is just a safer way to use T* in that you can increment: ++, decrement: -- but it is just better, there are set member functions get and get_val and not accidently using a pointer rather than the value conatained beneath the pointer. It also has the size member variable. pse::Iterator.begin() return the value datamember because that is the start. Also pse::Iterator.end() returns a default initialized pse::Iterator<T> like this return pse::Iterator<T>{};. We do this because we are using a higher level iterator and pse::Iterator<T>{} is a good sentinel value.

Also in the Algo.cpp header file, there are many algorithms:

template<typename Callable>
pse::Invoke(Callable& c)
{
  return c();
}

It is just like std::invoke but oviously in the pse library. It just returns the invoked callable passed in. It can be a lambda expression: [](){} a struct that has the operator() like this struct s{auto operator();} and that would work. It is a very simple function.

There is also pse::OnAll which takes an Iterator and a lamba and calls that lambda on all of the elemts:

template<typename T, typename Callable>
auto OnAll(Iterator<T>& it, Callable& c)
{
  // calls the lamdba on all of the it elements in the Iterator.
  // Only works if the constructor was passed a size.
}

pse::OnAll acts just like pse::for_each. There is no error handling for this though so if you pass in a lambda that takes an pse::StringList and you pass an Iterator<T> of type int you will just get an error. So use the power to pass in what you want wisely!

There is also pse::Sort it sorts all the elements in a pse::Iterator<T> by using the pse::swap overload for pse::Iterator. Its implementation looks like this:

template<typename T>
auto Sort(Iterator<T>& it)
{
  // sorts by increasing value.
}

But there is also an overload that takes a lambda/callable to use on pse::Sort. It looks like this:

template<typename T, typename Callable>
auto Sort(Iterator<T>& it)
{
  // sort based on what value the callable/lambda returns.
  // when using this function the lambda/callable has to return a bool value.
}

pse::pack_size is a way to find the size of a parameter pack at compile time. It works a bit like a pse::Tuple in that it stores it's elements like pse::Tuple<Ts...> other; but it is for finding the size of a parameter pack instead. Say you have a parameter pack of int, double, bool, char and we didn't know the size of this parameter pack, we could use pse::pack_size by doing pse::pack_size<int, double, bool, char>::value and that will give us the value of 4. An example implementation looks like this:

template<typename T, typename... Ts>
struct pack_size
{
};

template<typename T>
struct pack_size<T>
{
};

It is very simple, and all you have to do to access to size is to get the static datamember: value. There is also a container called pse::Bitmap, it consists of a collection of datamembers that are aranged like this:

	[0, 0], [1, 0], [2, 0], [3, 0], [4, 0], [5, 0]
	
	[0, 1], [1, 1], [2, 1], [3, 1], [4, 1], [5, 1]
	
	[0, 2], [1, 2], [2, 2], [3, 2], [4, 2], [5, 2]
	
	[0, 3], [1, 3], [2, 3], [3, 3], [4, 3], [5, 3]
	
	[0, 4], [1, 4], [2, 4], [3, 4], [4, 4], [5, 4]
	
	[0, 5], [1, 5], [2, 5], [3, 5], [4, 5], [5, 5]

pse::Bitmap's implementation looks a bit like this:

struct Bitmap
{
  // all of the usual constructors +
  template<typename T, typename... Ts>
  Bitmap(T t, Ts... ts);
  
  int m_bitmap[1][1];
};

It's constructor is very simple to use, you can simply pass in pse::Down{3}, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 and the pse::Bitmap will end up looking like this:

                 1, 6, 11
		 2, 7, 12
		 3, 8, 13
		 4, 9, 14
		 5, 10

We pass in pse::Down{} this tells us how many numbers we want across, in this case we passed in three, so it has three numbers down for example: 1, 2, 3, 4, 5 and the rest depends on how many numbers you passed in and the value of the value data member of pse::Down. The pse::Down shows us that we want pse::Down::value going down and the rest depends on what parameters we have passed in.

In the Traits.cpp header file, there are many useful classes/structs that are all just instances of meta-programming. In the traits library there is also an _v option for nearly every trait bar a few like add_ref_ref, and add_pointer_pointer. For example pse::Is_Type, it's implementation looks like this:

template<typename T>
struct Is_Type
{
};

pse::Is_Same compares two types to see if they are the same using partial template specialization. An example implementation looks like this:

template<typename A,  typename B>
struct Is_Same
{
};

There is also pse::Add_Const which just adds const to a type passed in, an example implementation looks like this:

template<typename T>
struct Add_Const
{
};

There is also a class/struct called pse::Remove_Const which just removes the const from a value (only if the type passed in is const). An example implementation looks like this:

template<typename T>
struct Remove_Const
{
};

There is also the same as for pse::Add_Const and pse::Remove_Const, there is pse::Add_Ref and pse::Remove_Ref. An example implementation of pse::Add_Ref looks like this:

template<typename T>
struct Add_Ref
{
};

And an example implemenatation of pse::Remove_Ref looks like this:

template<typename T>
struct Remove_Ref
{
};

There is also pse::Add_Pointer which adds a pointer to a given type. An example implementation of pse::Add_Pointer looks like this:

template<typename T>
struct Add_Pointer
{
};

Also you can check if a type is a pointer by doing this:

auto x = pse::Is_Same<pse::Add_Pointer<T>::value, T>::value;

To check wheather adding a pointer to an already pointer is the same type. Or you can just do:

pse::Is_Pointer<T>::value;

An example implementation of pse::Is_Pointer is:

template<typename T>
struct Is_Pointer
{
};

There is also pse::Remove_Volatile and with it comes pse::Remove_Volatile_v. Given a type it removes the volatile from it (if there is any). An example implementation looks like this:

template<typename T>
struct Remove_Volatile
{
  // Implementation.
};

template<typename T>
struct Remove_Volatile_v = Remove_Volatile<T>::value;

There is also remove_const_volatile_ref which removes all type specifiers. For example if it is given const volatile& char it's value would be char. It's implementation looks like this:

template<typename T>
struct remove_const_volatile
{
    using value = Remove_Const_v<Remove_Volatile_v<Remove_Ref_v<T>>>;
};

There is also remove_const_volatile_ref_v which is just like Remove_Const_v and Remove_Volatile_v in that it gets the value out of the class remove_const_volatile_ref with the type T. An example implementation looks like this:

template<typename T>
using remove_const_volatile_ref_v = remove_const_volatile_ref<T>::value;

Also there is pse::GeneratorSame which is just a specialised pse::Generator for a gererator of the same type. It still has the same interface as pse::Generator but just small changes in the internal mechanisms. An example implementation looks like this:

template<auto T, auto... Ts>
struct GeneratorSame
{
  TupleCTAD<T, Ts...> types{};
  auto operator()();
};

So, as stated before the interface is just like the one seen in pse::Generator, so for more information look at the details of the pse::Generator implemenatation. The only difference between the two is that pse::GeneratorSame has a different internal mechanism of operator()().

There are also lots of other algorithms in the Algo.cpp header file. One example is pse::find it just takes a data structure that has the operator[] and takes something to find. If it finds that thing in the data structure it returns the index that it was found at, if nothing was found then 0 will be retuned. An example implementation of pse::find looks like this:

auto find(auto& a, auto& ToFind)
{
 // Implemenatation.
}

There is also pse::find_if which finds an element in a data structure if it meets the requirement of the callable passed in. An example implementation looks like this:

auto find_if(auto& a, auto& Callable)
{
  // Implemenatation.
}

If the callable returns true with a given index of the data structure, that index is returned otherwise 0 is returned, to symbol that no value that met the condition of the Callable(something with operator()) evaluated to true in the data structure.

There is also pse::reverse which takes a given container a and reverses the elements in that array. The implementation looks a bit like this:

auto reverse(auto& a)
{
  // reverse the elements in the array.
}

So having a given input: 1, 2, 3, 4, when passed to the reverse function, pse::reverse it will now look like: 4, 3, 2, 1. Which is succesfully reversed! But the container that stores the elements it has to have operator[] and the types stored have to support operator= and those are mostly all of the constraints!

There is also pse::apply what this does it takes a range r and then takes two operations, minus and times for example. Say the range, r was 1, 2, 3, 4 and the operands were pse::minus and pse::multiply. It would look like this:

 (((0 -= (1 * 2)) -= (2 * 3)) -= (3 * 4))

An example implementation of pse::apply looks like this:

template<typename TypeOfRange>
auto apply(TypeOfRange& r, auto& first_operand, auto& second_operand)
{
  // implementation.
}

There is also pse::abs, this function gives the absolute value of a number. This is the distance between 0 and the number, with no regard to direction. So if i has a number -5 and i used pse::abs the output would be 5 because the distance between -5 and 0 is 5. So if i passed in 5 to pse::abs it would give me the same answer 5. It is basically just a way of getting rid of zero. Here is an example implementation:

inline double abs(double i)
{
	if (i < 0)
	{
		return (i + (0 - i) + i); // this bit here is the magic.
	}
	else
	{
		return i;
	}
}

There is also pse::floor this just rounds down to the nearest whole number, an example implemenation looks like this:

auto abs(double b)
{
  // Implementation.
}

There is also pse::ciel which is just basically the opposite of pse::floor in that is rounds up to the nearest whole number. An example implemenatation looks like this:

auto ciel(double b)
{
  // Implemenatation.
}

There is also pse::round which rounds to the nearest whole number. So if x < 0.5 round down otherwise, x => 0.5 round up. An example implemenation looks like this:

auto round(double b)
{
  // Implemenatation.
}

There is also the famous gcd algorithm that find the Greatest Common Divisor. An example implemenatation looks like this:

int gcd(int a, int b) {
  // base case
  if (a == 0) {
    return b;
  }
  if (b == 0) {
    return a;
  }

  // recursive case
  if (a > b) {
    return gcd(a % b, b);
  } else {
    return gcd(a, b % a);
  }
}

There is also the LCM algorithm, the Lowest Common Multiple algorithm finds the lowest common multiple of two numbers, an example implementation looks like this:

int lcm(int a, int b)
{
    // Find the greatest common divisor of a and b
    int gcd = 1;
    for (int i = 2; i <= a && i <= b; i++)
    {
        if (a % i == 0 && b % i == 0)
            gcd = i;
    }

    // Return the LCM of a and b
    return (a * b) / gcd;
}

There is also pse::Node this is just like any other node that you would find, it has a parent: parent and two child nodes: left_child and right_child. An example implemenation of pse::Node looks like this:

template<typename T>
struct Node
{
  T m_data;

  Node<T>* parent;
  Node<T>* first_child;
  Node<T>* second_child;
};

There is also a pse::BinaryTree, it looks something like this:

template <typename T>  // template for type of Node.
struct BinaryTree {
    Node<T>* curr;

    auto left() {
        if (curr->first_child != nullptr) curr = curr->first_child;
    }
    auto right() {
        if (curr->right_child != nullptr) curr = curr->second_child;
    }
    auto up() {
        if (curr->parent != nullptr) curr = curr->parent;
    }

    auto get() { return curr.m_data; }

    template <typename NodeType, typename... NodeTypes>
    BinaryTree(NodeType* n, NodeTypes*... ns);

    template <typename NodeType>
    BinaryTree(NodeType* n);
};

But it is only partially implemented, so it is not production code ready yet. But it has quite a lot of the functionality, but the constructors are not the best.

There is also pse::enable_if, that is just like std::enable_if. It rely's on SFINAE (Substitution Failiure Is Not An Error). pse::enable_if can be replaced with other things like c++20 concepts, but in case you don't have a modern c++20 compiler, enable_if is implemented. It's implementation looks like this:

// the base template
template<bool B, typename T>
struct enable_if
{
};

// the specialisation
template<typename T>
struct enable_if<true, T>
{
  using value = T;
}

template<bool b, typename T>
using enable_if_v = enable_if<b, T>::value;

There are also several useful factory functions in the pse library, although they have obsure names they are very useful! There is factory_parameter_pack_non_typename there is factory_parameter_pack_typename, there is factory_construct_typename, there is factory_construct_non_typename.

Here are example inmplementations of them all:

factory_constructor_non_typename:

template<typename Class, detail::constructor& p_p, auto&&... Args>

	struct factory_constructor_non_typename
	{
		auto operator()()
		{
			return Value;
		}

		Class Value{ Args... };
	};
	

factory_parameter_pack_non_typename:

	
	template<typename Class, detail::parameter_pack& c, auto&&... Args>
	struct factory_parameter_pack_non_typename
	{
		auto operator()()
		{
			return Value;
		}

		Class Value = Class{ Args... };
	};

factory_construct_typename:

	template<typename Class, detail::constructor& c, typename... Args>
	struct factory_construct_typename
	{
		auto operator()(Args... args)
		{
			return Class{ args... };
		}
	};

factory_parameter_pack_typename:

	template<typename Class, detail::parameter_pack& p_p, typename... Args>
	struct factory_parameter_pack_typename
	{
		constexpr auto operator()(Args... args)
		{
			return Class{args...};
		}
	};
	

There is also char_int so that you can convert between chars and ints, it only works with single digit digits though because, for instance you cannot do this 432 -> '432' because that does not fit inside a character! There is int_to_char and char_to_int. Here is an example implementation of int_to_char:

  auto int_to_char(int i)
  {
    return (static_cast<char>(i + 48));
  }

And an example implementation of char_to_int is as follows:

  auto int_to_char(int i)
  {
    return (static_cast<char>(i + 48));
  }

pse::CodeMeaning contained in the CodeMeaning.cpp header file has lots of different if's and else if's in order to find the write error code, here is a collection of the error codes and their values.

There is also pse::index_sequence. This generates a sequence of indixes for the passed parameter pack T and Ts.... It will have a constant array indeces containing all of the indexes for the types that you have passed. An examle of calling it would look like this:

index_sequence<0, int, float, char, bool> i_s{};

so then the array indeces will have the following values: 0, 1, 2, 3. An example implementation of pse::index_sequence is as follows:

template<auto StartSize = 0, typename T, typename... Ts>
struct index_sequence
{
    decltype(StartSize) indeces[sizeof...(Ts) + 1){/*indeces*/};
    
    // rest of implementation.
};
026: Out Of Bounds Access.
034: Your hardware is not supported.
054: Unvalid Type.
056: Boolean evaluated to false.
076: Not A Number (NaN).
84: This Function Is No Longer In Use.
96: Your Software is not supported.
Some Useful Links

The pse Library blog! https://rshepherdcpp.wixsite.com/none/post/the-pse-library-new-idiom

SonarLink: https://sonarcloud.io/project/configuration?id=robertshepherdcpp_pse&analysisMode=GitHubActions