A minimal but powerful programming language based on C, Go and Haskell
Because a lot of non-essential features are removed from the language and core, the user has freedom to implement them however they want. In this section we provide some of possible solutions for these types of challenges.
Polymorphism can be achieved using unions, type casting and error handling.
We need corresponding functions to accept nothing as an input. In this case, caller will cast union to the correct type, or nothing if not possible.
drawCircle = fn(x: Circle, c: Canvas, f: Color->int) {...}
drawSquare = ...
drawTriangle = ...
superDraw = fn(s: Shape->nothing) {
#here key is type, and value is a lambda which takes a canvas and a color and returns error or int
draw = [
Circle: fn{ drawCircle(getType(x, Circle)@,_,_) },
Square: fn{ drawSquare(getType(x, Square)@,_,_) },
Triangle: fn{ drawTriangle(getType(x, Triangle)@,_,_) }
]
drawFunction = draw[type(s)]() #get compile time type of "s"
drawFunction()
}
If you want to add a new shape (e.g. Oval), you should update the Shape type definition and add appropriate draw functions and update the superDraw function.
If you want to add a new operation (e.g. print), you will need to add a new function (e.g. superPrint).
It is advised to put all import dependency paths in one module like refs and import it to specify import paths.
#refs
std_map = "/http/github.com/dotLang/std/v1.9.5/MapHelper"
and then use above to do actual import:
#File1 module
refs = import("/src/main")
MapHelper = import(refs..std_map)
Sometimes we need to have a generic function that can accept any other function (regardless of input count/type and output type). This is not directly possible in dotLang but there are other alternatives that can help in these cases. For example suppose that you have a function that accepts an optional validation function. If the function is provided, you want to call it, but if not, you want to ignore. You can define a generic function that can do this as here, we only care about the output of the validation function:
NopFunc1 = fn(input: T, T: type -> nothing)
NopFunc2 = fn(input1: T, input2: U, T: type, U: type -> nothing)
NopFunc3 = fn(input1: T, input2: U, input3: V, T: type, U: type, V: type -> nothing)
...
result = (validator//NopFunc3)(x, y, z)
Many languages have switch, case or match keyword to do pattern matching. in dotLang we use hashmaps as a simple alternative.
int_result =
[
value1: fn{1},
value2: fn{2}
][exp]()
main = fn( -> int ) { 0 }
This is a function, called main which has no input and always returns 0 (very similar to C/C++ except main function has no input).
You can simplify main to:
main = fn{ 0 }
main = fn( -> int)
{
print("Hello world!")
0
}
We want to write a function which accepts a string like "2+4-3" and returns the result (3).
NormalExpression = struct (op: char, left: Expression, right: Expression)
Expression = int|NormalExpression
eval = fn(input: string -> float)
{
#assume we already have parse function
exp = parse(input)
innerEval(exp)
}
innerEval = fn(exp: Expression -> float)
{
int_val, normal_exp = exp
int_val // fn{
#now we are sure that exp is of type NormalExpression
y = unwrap(normal_exp)
['+' : fn{innerEval(y.left) + innerEval(y.right)},
'-' : fn{innerEval(y.left) - innerEval(y.right)},
'*' : fn{innerEval(y.left) * innerEval(y.right)},
'/' : fn{innerEval(y.left) / innerEval(y.right)}][y.op]()
}()
}
quickSort = fn(list:[int], low: int, high: int -> [int])
{
ifElse(high<low, list, fn{
mid_index = (high+low)/2
pivot = list[mid_index]
#filter is a built-in function
small_list = filter(list, fn(x:int -> bool) {x < pivot}))
big_list = filter(list, fn(x:int -> bool) {x > pivot}))
quickSort(small_list) + [pivot] + quickSort(big_list)
})
}
A function which accepts a list of numbers and returns sum of numbers.
sum = (data: [int] -> int)
{
calc = (index: int, sum: int -> int)
{
iElse(index >= length(data), fn{sum}
fn{calc(index+1, sum + data[index]})()
}
calc(0,0)
}
A function which accepts a number and returns its digits in a sequence of characters. Generally for this purpose, using a linked-list is better because it will provide better performance, but we are using recursion to show how it works.
extractor = (n: number, result: string -> string)
{
ifElse(n < 10, fn{result + char(48+n)}, fn{
digit = n % 10
extractor(n/10, result + char(48+digit)
})
}
A function which accepts two sequences of numbers and returns the maximum of sum of any any two numbers chosen from each of them. This can be done by finding maximum element in each of the arrays but we want to do it with a nested loop.
maxSum = (a: [int], b: [int] -> int)
{
calc = (idx1: int, idx2: int, current_max: int -> int)
{
ifElse(idx2 >= length(b), fn{current_max}, fn{
sum = a[idx1] + b[idx2]
next1 = (idx1+1) % length(a)
next2 = idx2 + int((idx1+1)/length(a))
calc(next1, next2, max(current_max, sum))
})
}
calc(0, 0, 0)
}
fib = (n: int, cache: [int|nothing] -> int)
{
ifElse(cache[n] != nothing, unwrap(cache[n], int),
fn{
seq_final1 = set(seq, n-1, fib(n-1, cache))
seq_final2 = set(seq_final1, n-2, fib(n-2, seq_final1))
seq_final2[n-1] + seq_final2[n-2]
})
}
std = import("/http/github.com/dotLang/std")
stdin = std..stdin
stdout = std..stdout
rand = std..random
main = fn
{
secret = rand(1,100)
stdout.write("Please enter a number: ");
guessRaw: int|nothing = tryParseString(int, stdin.readLine())
ifElse(guessRaw == nothing,
fn{
stdout.write("Invalid number")
},
fn {
guess = unwrap(guessRaw)
actions = [guess<secret: fn{ stdout.write("Too small!") },
guess>secret: fn{ stdout.write("Too large!") },
guess==secret: fn{ stdout.write("Well done!") }
][true]()
})
}