GitHub

SugarPP: syntactic 🍬 for programming in C++

Build

SugarPP is a collection of syntactic sugar for C++ code.

How to Use

  1. SugarPP is header only and each header file is independent. Just clone this repository and go to ./include/sugarpp or copy the corresponding header file you want to use.

    Or, if you want to use it hassle free, copy this Cmake snippet to your root CMakeLists.txt to automatically download & use this library in your project. No need to clone the project manually!

    include(FetchContent)
FetchContent_Declare(
    SugarPP
    GIT_REPOSITORY https://github.com/HO-COOH/SugarPP.git
    GIT_TAG origin/master
)
FetchContent_MakeAvailable(SugarPP)
#Use for your target
add_executable(<Your target> main.cpp)
target_link_libraries(<Your target> PRIVATE SugarPP)

    Or, if you want to use the library globally, copy this Cmake snippet to your roor CMakeLists.txt to automatically download & use this library in your project. No need to clone the project!

    include(FetchContent)
FetchContent_Declare(
    SugarPP
    GIT_REPOSITORY https://github.com/HO-COOH/SugarPP.git
    GIT_TAG origin/master
)
FetchContent_MakeAvailable(SugarPP)
#Use globally
link_libraries(SugarPP)

  2. Then add #include <sugarpp/xxx/xxx.hpp>. Also Note: Everything in SugarPP is now inside namespace SugarPP! So you may want using namespace SugarPP;

You can find quick documentation for every modules in ./docs

You can find examples for every modules in ./test/source

Alternatively, see generated doxygen document here.

Requirements

SugarPP uses various C++17 language features; thus, it requires a C++17 compatible compiler to use.

GCC 10.1 and older has a known bug, which causes issues on the overload resolution of detail::when_impl; consider upgrading to GCC 10.2 or newer Nope, it is compatible with GCC 9.2 now :)

Tested with:

  • GCC 10.2 & GCC 9.2
  • Clang 10.0
  • Visual Studio 16.7

Features

Kotlin-style when syntax

Kotlin has the when expression for matching values, replacing the traditional switch/case in most languages. C/C++'s native switch/case has the drawback of only matching integer values.

SugarPP when

  • Value matching (for any comparable type, not just integers):
/* Kotlin */
when (x) {
    1 -> print("x == 1")
    2 -> print("x == 2")
    else -> { // Note the block here
        print("x is neither 1 nor 2")
    }
}
/* SugarPP */
when (x,
    1,      []{ print("x == 1"); },
    2,      []{ print("x == 2"); },
    Else(), []{ print("x is neither 1 nor 2"); }
)(); //returns a function object, use () to call it
  • Type matching:
/*kotlin*/
fun describe(obj: Any): String =
    when (obj) {
        1          -> "One"
        "Hello"    -> "Greeting"
        is Long    -> "Long"
        !is String -> "Not a string"
        else       -> "Unknown"
    }
/*SugarPP*/
auto describe = [](auto&& obj) {
    return when(obj,
        1,                      "One",
        "hello",                "Greeting",
        is<long>(),             "long",
        is_not<const char*>(),  "Not a string",
        Else(),                 "Unknown string"
    );
};
  • Polymorphic type matching:
/*SugarPP*/
struct Shape { virtual ~Shape() = default; };
struct Circle :Shape {};
struct Square :Shape{};
std::unique_ptr<Shape> pt{ new Circle{} };
when(*pt,
    is_actually<Circle>(), [] { print("Circle* pt"); },
    is_actually<Square>(), [] { print("Square* pt"); },
    Else(),                [] { print("Unknown type"); }
)();    //"Circle* pt"
  • Range matching:
/*kotlin*/
val validNumbers = arrayOf(11, 13, 17, 19)
when (x) {
    in 1..10 ->         print("x is in the range")
    in validNumbers ->  print("x is valid")
    !in 10..20 ->       print("x is outside the range")
    else ->             print("none of the above")
}
/*SugarPP*/
std::array validNumbers{11,13,17,19};
when(x,
    Range(1, 10),       []{ print("x is in the range"); },
    Range(validNumbers),[]{ print("x is valid"); },
    NOT{Range(10, 20)}, []{ print("x is outside the range"); },
    Else(),             []{ print("none of the above"); }
)();
  • Argument-less switches
/*kotlin*/
when {
    x.isOdd() ->    print("x is odd")
    y.isEven() ->   print("y is even")
    else ->         print("x+y is even.")
}
/*SugarPP*/
int x = 1, y = 2;
when(
    isOdd(x),   []{ print("x is odd"); },
    isEven(y),  []{ print("y is even"); },
    Else(),     []{ print("x+y is even");}
)();//"x is odd"

Note: Unlike C/C++ switch-case, kotlin when is short-circuiting; the execution terminates at the first satisfied branch. SugarPP when has the same behavior.

  • Pattern matching

Kotlin doesn't seems to support _ as a place holder.

/*Rust*/
for i in 1..=100 {
    match (i % 3, i % 5) {
        (0, 0) => println!("FizzBuzz"),
        (0, _) => println!("Fizz"),
        (_, 0) => println!("Buzz"),
        (_, _) => println!("{}", i),
    }
}
/*SugarPP*/
for(auto i:Range(1, 101))
{
    when(std::tuple{ i % 3, i % 5 },
        std::tuple{ 0, 0 }, [] { print("fizzbuzz"); },
        std::tuple{ 0, _ }, [] { print("fizz"); },
        std::tuple{ _, 0 }, [] { print("buzz"); },
        Else(),             [i] { print(i); }
    )();
}

Usage

Just copy ./include/sugarpp/when/when.hpp and add #include "when.hpp" in your project.

Documentation

See docs/When.md

Read More

At the time of writing this library, I was not aware of the C++23 pattern matching proposal. And yes, SugarPP::when will have performance penality compared with what can be done with switch-case statement. SugarPP::when works as recursively comparing the condition to each branch, so I am not sure whether this has performance penalty compared with the pattern matching proposal.

As I am still an early learner, I will update this part to give you more insight. You can find the original implementation of that proposal here

Simpler IO

IO in C++ should work how you expect it to. SugarPP's IO functions are simple and much more intuitive than native C++ IO. No more messing with getchar() and getline() nonsense, and print() anything!

Features

The input template function is similar to Python's input. It prints a prompt message and does automatic error handling - for example, if the input is bad, it will clear the bad bit and re-prompt until an acceptable input is given (this behavior can be disabled).

  • If the type is a primitive, the function will work the same as std::cin >>
    • If the type given is unsigned, input will automatically convert the input to an absolute value.
  • If the type is std::string, it will behave the same as std::getline, getting the whole line at once.
auto name = input<std::string>("Enter your name: ");
print("Hello,", name, "How old are you?");
auto age = input<int>("Enter your age: ");
print(name, "is", age, "years old");

The print function also behaves similar to Python's print; it can print any number of arguments of any type, separated by a specified delimiter (defaulting to space). SugarPP's print can print almost anything:

  • Anything std::cout has an overload for
  • Anything that is iterable (i.e. has a .begin() or can be called with std::begin)
  • Nested iterables (at any depth)
  • std::tuple and std::pair
  • bool will be printed as True or False

printLn behaves similarly, but prints each argument on a new line.

/*print any iterable*/
std::array arr{ 1,2,3 };
print(arr); //[1, 2, 3]
/*print a tuple*/
std::tuple t{ "SugarPP", 123, 45.6f };
print(t); //(SugarPP, 123, 45.6)
/*print any nested printable*/
std::vector<std::vector<int>> v1{ {1,2,3}, {5,6,7,8}, {9,10} };
std::vector<std::vector<std::vector<int>>> v2{ {{1,2,3}, {5,6,7,8}, {9,10}}, {{10,11},{12,13}, {}} };
printLn(v1, v2); //[[1, 2, 3], [5, 6, 7, 8], [9, 10]]...
/*print a bool*/
print(0.1 + 0.2 == 0.3); //"False", you should know why :P

There are additional ThreadSafe versions of these functions with the same name, under namespace ThreadSafe.

Usage

Just copy ./include/sugarpp/io/io.hp and add #include "io.hpp".

More examples in ./test/source/io/io.cpp.

Documentation

See docs/IO.md.

Range

Use numerical ranges to simplify your range-based for loop! Not to be confused with C++20 ranges. Container Ranges are working in progress towards providing C++20 ranges functionality in C++17.

Many other programming languages have a range syntax for iteration:

// e.g. in Rust
for n in 0..10 {
    println!(n);
}
# or in Python
for i in range(0, 10):
    print(i)

Features

SugarPP defines 3 types of Ranges in some sort of "class overloading" way

  • Numerical ranges

    • start, end, and step (default = 1) with a C++ foreach loop. Type will be inferred and automatically converted if needed. 
      for (auto i : Range(2.0, 10.0, 3))
    print(i);
/*
    2
    5
    8
*/
    • Multiple-dimension ranges 
      for (auto [i, j] : Range(-5, 1) | Range(0, 3))
    print(i, '\t', j);
/*
    -5       0
    -5       1
    -5       2
    -4       0
    -4       1
    -4       2
    ...
    0        0
    0        1
    0        2
*/
    • Generating a random number within the range 
      /*use range for a random number*/
Range r(-1, 100000);
print("Random number in ", r, " is ", r.rand());
/*use range to generate several random numbers*/
auto [num1, num2, num3] = Range(1, 10).rand<3>();
    • Filling a container with random numbers 
      /*use range to fill a C style array*/
double arr[10];
Range(-500.0, 500.0).fillRand(arr);
/*use range to fill a STL container*/
std::array<char, 20> arr2;
Range('A', 'z').fillRand(arr2);
/*Alternatively .randFast() provides a faster way for generating random number using rand() in C*/
int arr3[10];
Range(-200, 300).fillRandFast(arr3);

  • Letter ranges

    Similar functionality with numerical ranges, but works correctly when it is incremented it skips non letter characters

  • Container ranges(In progress)

SugarPP also has an Enumerate class, which returns a pair of index (default to start at 0) and a reference to the content of iterable, similar to Python's enumerate().

# Python
a = ["enumerate", "in", "python"]
for i, content in enumerate(a):
    print(i, content)
/*SugarPP*/
std::array a{"Enumerate", "in", "SugarPP"};
for(auto [i, content] : Enumerate(a))
    print(i, content);

Usage

Just copy ./include/sugarpp/range/range.hpp and add #include "range.hpp" for Range.

Just copy ./include/sugarpp/range/enumerate.hpp and add #include "enumerate.hpp" for Enumerate.

More examples in ./test/source/range/range.cpp

Documentation

See docs/Range.md.

Type conversion

Features

  • To & from string

    • -> number

      Admit it, atoi(), atof(), wcstold(), strtof() are some of the ugliest function name in C, and C++ makes it worse by adding more obsecure names like std::stoi(), std::stolld(). That's why SugarPP provides a uniform way of getting numbers from string, which is SugarPP::to_num<Type>(), which accepts both normal strings and wide-strings.

      /*SugarPP*/
auto str1 = "42";
auto num1 = to_num<int>(str1);
auto str2 = "3.14159";
auto num2 = to_num<double>(str2);

    • -> string

      Isn't it wired that something printable can't be converted to string? Isn't it wired that there is a std::to_string() only works for numerical values? SugarPP::to_string not only works with anything that can be converted with std::to_string(), but also anything that is printable. Additionally, you can also specify whether it's normal character or wide-character using one template argument.

      auto f_str = to_string(23.43);
std::ostream& operator<<(std::ostream& os, const MyPrintableClass&);
auto my_class_str = to_string(my_printable);

Usage

Just copy ./include/sugarpp/types/types.hpp and add #include "types.hpp".

More example in ./test/source/types/to_string.cpp

Lazy

Features

A C++ implementation for Kotlin's Lazy, which represents a value with lazy initialization, with type inference from the initializer and multi-threading synchronization support.

Lazy lazyInt{ [] { return 1; } }; //Lazy<Int>
Lazy lazyDouble
{
	[] {
        /*Some computation*/
		return 2.0;
	},
	ThreadSafetyMode::Synchronized
};

Possible thread-safety modes are also equivalent to Kotlin

enum class ThreadSafetyMode
{
	Synchronized,
	Publication,
	None
};

Motivation

I had so much fun writing these and learned so much. Such a great language that gives you nightmare everytime you want to add stuff. C++ itself is difficult enough, yet you realize that you can't even have a compiler to trust with when 3 different compilers (Visual studio, Clang, GCC) gives you different results.

Read the original on github.com ↗