MonkeyLang

Monkey Programming Language written in Julia.

Table of Contents

• MonkeyLang
  • Using Monkey in Julia
  • Compile MonkeyLang.jl to a standalone executable
  • Start the REPL
  • Documentation
    • Summary
    • Syntax overview
      • If
      • While
      • Operators
      • Return
    • Variable bindings
    • Scopes
      • Global Scope
      • Local Scope
      • Closure Scope
      • CurrentClosure Scope
    • Literals
      • INTEGER
      • BOOLEAN
      • NULL
      • STRING
      • ARRAY
      • HASH
      • FUNCTION
    • Built-in Functions
      • type(<arg1>): STRING
      • puts(<arg1>, <arg2>, ...): NULL
      • len(<arg>): INTEGER
      • first(<arg: STRING>): STRING | NULL
      • first(<arg: Array>): any
      • last(<arg: String>): STRING | NULL
      • last(<arg: Array>): any
      • rest(<arg: STRING>): STRING | NULL
      • rest(<arg: ARRAY>): ARRAY | NULL
      • push(<arg1: ARRAY>, <arg2>): ARRAY
      • push(<arg1: HASH>, <arg2>, <arg3>): HASH
    • Advanced examples
      • A custom map function
      • A custom reduce function
    • Macro System

Using Monkey in Julia

You can start the REPL within Julia:

using MonkeyLang

start_repl()

Or you can evaluate Monkey programs using string macros:

using MonkeyLang

a = 2

monkey_eval"let b = $a; puts(b)"

monkey_vm"let c = [$a, $a]; puts(c)"

monkey_julia"let d = {$a: $a}; puts(d)"

Compile MonkeyLang.jl to a standalone executable

Clone the repo, and run make build in the root directory.

Caution: The compilation may take up to ~5 minutes.

Start the REPL

You can start the REPL in a Julia script or in the Julia REPL:

import Pkg; Pkg.add("MonkeyLang")

using MonkeyLang

MonkeyLang.start_repl()
MonkeyLang.start_repl(; use_vm = true) # use VM

You can press Ctrl-C or Ctrl-D to exit the REPL.

If you have compiled MonkeyLang.jl locally, then you can directly start the REPL by:

./monkey repl
./monkey repl --vm # use VM

Documentation

I created the document with reference to Writing An Interpreter In Go and rs-monkey-lang.

⚠️ Please note that there may be some mistakes.

Summary

• C-like syntax
• variable bindings
• first-class and higher-order functions • closures
• arithmetic expressions
• built-in functions

Syntax overview

An example of Fibonacci function.

let fibonacci = fn(x) {
  if (x == 0) {
    0;
  } else {
    if (x == 1) {
      1;
    } else {
      fibonacci(x - 1) + fibonacci(x - 2);
    }
  }
};

fibonacci(10);

If

It supports the general if. else exists, but else if does not exist.

if (true) {
  10;
} else {
  5;
}

While

It also supports while loops.

let x = 5;
while (x > 0) {
  puts(x);
  x = x - 1;
}

Operators

It supports the general operations.

1 + 2 + (3 * 4) - (10 / 5);
!true;
!false;
+10;
-5;
"Hello" + " " + "World";

Return

It returns the value immediately. No further processing will be executed.

if (true) {
  return;
}

let identity = fn(x) {
  return x;
};

identity("Monkey");

Variable bindings

Variable bindings, such as those supported by many programming languages, are implemented. Variables can be defined using the let keyword. Variables cannot be redefined in the same scope, but they can be reassigned.

Format:

let <identifier> = <expression>; # Define

<identifier> = <expression>; # Reassign

Example:

let x = 0;
let y = 10;
let foobar = add(5, 5);
let alias = foobar;
let identity = fn(x) { x };

x = x + 1;
y = x - y;

Scopes

In Monkey, there are types of scopes:

Global Scope

Variables defined at the top level are visible everywhere, and can be modified from anywhere.

let x = 2; # `x` is a global variable

let f = fn() { 
  let g = fn() { 
    x = x + 1; # Modifies the global variable `x`
    return x; 
  } 
  return g; 
}

let g = f();
puts(g()); # 3
puts(g()); # 4

let h = f();
puts(h()); # 5
puts(h()); # 6

Local Scope

Variables defined within while loops or functions are of this scope. They can be modified from the same scope, or inner while loops' scopes.

let x = 1;

while (x > 0) {
  x = x - 1;
  let y = 1; # `y` is a local variable
  while (y > 0) {
    y = y - 1; # Modifies the local variable `y`
  }
  puts(y); # 0
}

Closure Scope

A function captures all non-global variables visible to it as its free variables. These variables can be modified from within the function.

let f = fn() { 
  let x = 2; 
  let g = fn() { 
    x = x + 1; # `x` is captured as a free variable
    return x; 
  } 
  return g; 
}

let g = f();
puts(g()); # 3
puts(g()); # 4

let h = f();
puts(h()); # 3, since in function `f`, `x` remains unchanged.
puts(h()); # 4

CurrentClosure Scope

Specially, a named function being defined is of this scope. It cannot be modified from within its body.

let f = fn(x) {
  f = 3; # ERROR: cannot reassign the current function being defined: f
}

But redefinition is OK:

let f = fn(x) {
  let f = x + x;
  puts(f);
}

f(3); # 6

Literals

Five types of literals are implemented.

INTEGER

INTEGER represents an integer value. Floating point numbers can not be handled.

Format:

[-+]?[1-9][0-9]*;

Example:

10;
1234;

BOOLEAN

BOOLEAN represents a boolean value.

Format:

true | false;

Example:

true;
false;

let truthy = !false;
let falsy = !true;

NULL

NULL represents null. When used as a condition, NULL is evaluated as falsy.

Format:

null;

Example:

if (null) { 2 } else { 3 }; # 3

STRING

STRING represents a string. Only double quotes can be used.

STRINGs can be concatenated with "+".

Format:

"<value>";

Example:

"Monkey Programming Language"; # "Monkey Programming Language";
"Hello" + " " + "World"; # "Hello World"

ARRAY

ARRAY represents an ordered contiguous element. Each element can contain different data types.

Format:

[<expression>, <expression>, ...];

Example:

[1, 2, 3 + 3, fn(x) { x }, add(2, 2), true];

let arr = [1, true, fn(x) { x }];

arr[0];
arr[1];
arr[2](10);
arr[1 + 1](10);

HASH

HASH expresses data associating keys with values.

Format:

{ <expression>: <expression>, <expression>: <expression>, ... };

Example:

let hash = {
  "name": "Jimmy",
  "age": 72,
  true: "a boolean",
  99: "an integer"
};

hash["name"];
hash["a" + "ge"];
hash[true];
hash[99];
hash[100 - 1];

FUNCTION

FUNCTION supports functions like those supported by other programming languages.

Format:

fn (<parameter one>, <parameter two>, ...) { <block statement> };

Example:

let add = fn(x, y) {
  return x + y;
};

add(10, 20);

let add = fn(x, y) {
  x + y;
};

add(10, 20);

If return does not exist, it returns the result of the last evaluated expression.

let addThree = fn(x) { x + 3 };
let callTwoTimes = fn(x, f) { f(f(x)) };

callTwoTimes(3, addThree);

Passing around functions, higher-order functions and closures will also work.

The evaluation order of function parameters is left to right.

So a memoized Fibonacci function should be implemented like:

let d = {}

let fibonacci = fn(x) {
    if (x == 0) {
        0
    } else {
        if (x == 1) {
            1;
        } else {
            if (type(d[x]) == "NULL") {
                # You cannot use `d = push(d, x, fibonacci(x - 1) + fibonacci(x - 2))`
                # since `d` is evaluated first, which means it will not be updated
                # when `fibonacci(x - 1)` and `fibonacci(x - 2)` are called.
                let g = fibonacci(x - 1) + fibonacci(x - 2);
                d = push(d, x, g);
            }

            d[x];
        }
    }
};

fibonacci(35);

Built-in Functions

You can use the following built-in functions 🚀

type(<arg1>): STRING

Return the type of arg1 as a STRING.

type(1); # INTEGER
type("123"); # STRING
type(false); # BOOLEAN

puts(<arg1>, <arg2>, ...): NULL

It outputs the specified value to stdout. In the case of Playground, it is output to console.

puts("Hello");
puts("World!");

len(<arg>): INTEGER

• For STRING, it returns the number of characters.
• For ARRAY, it returns the number of elements.
• For HASH, it returns the number of key-value pairs.

len("Monkey"); # 6
len([0, 1, 2]); # 3
len({1: 2, 2: 3}); # 2

first(<arg: STRING>): STRING | NULL

Returns the character at the beginning of a STRING. If the STRING is empty, return NULL instead.

first("123"); # "1"
first(""); # null

first(<arg: Array>): any

Returns the element at the beginning of an ARRAY. If the ARRAY is empty, return NULL instead.

first([0, 1, 2]); # 0
first([]); # null

last(<arg: String>): STRING | NULL

Returns the element at the last of a STRING. If the STRING is empty, return NULL instead.

last("123"); # "3"
last(""); # null

last(<arg: Array>): any

Returns the element at the last of an ARRAY. If the ARRAY is empty, return NULL instead.

last([0, 1, 2]); # 2
last([]) # null

rest(<arg: STRING>): STRING | NULL

Returns a new STRING with the first element removed. If the STRING is empty, return Null instead.

rest("123"); # "23"
rest(""); # null

rest(<arg: ARRAY>): ARRAY | NULL

Returns a new ARRAY with the first element removed. If the ARRAY is empty, return NULL instead.

rest([0, 1, 2]); # [1, 2]
rest([]); # null

push(<arg1: ARRAY>, <arg2>): ARRAY

Returns a new ARRAY with the element specified at the end added.

push([0, 1], 2); # [0, 1, 2]

push(<arg1: HASH>, <arg2>, <arg3>): HASH

Returns a new HASH with arg2: arg3 added. If arg2 already exists, the value will be updated with arg3.

push({0: 1}, 1, 2); # {1:2, 0:1}
push({0: 1}, 0, 3); # {0:3}

Advanced examples

A custom map function

let map = fn(arr, f) {
  let iter = fn(arr, accumulated) { 
    if (len(arr) == 0) {  
      accumulated 
    } else { 
      iter(rest(arr), push(accumulated, f(first(arr)))); 
    } 
  };

  iter(arr, []);
};

let a = [1, 2, 3, 4];
let double = fn(x) { x * 2};
map(a, double); # [2, 4, 6, 8]

A custom reduce function

let reduce = fn(arr, initial, f) {
  let iter = fn(arr, result) {
    if (len(arr) == 0) {
      result
    } else { 
      iter(rest(arr), f(result, first(arr)))
    }
  }

  iter(arr, initial)
}

let sum = fn(arr) { 
  reduce(arr, 0, fn(initial, el) { initial + el })
}

sum([1, 2, 3, 4, 5]); # 15

Macro System

Now that the Lost Chapter has been implemented, MonkeyLang.jl provides a powerful macro system.

Here is an example:

let unless = macro(condition, consequence, alternative) {
    quote(if (!(unquote(condition))) {
        unquote(consequence);
    } else {
        unquote(alternative);
    });
};

unless(10 > 5, puts("not greater"), puts("greater")); # greater

In the REPL, you need to enter all the contents in a single line without \n characters.

Enjoy Monkey 🐵 !