Wc in D: 712 Characters Without a Single Branch

After reading “Beating C With 80 Lines Of Haskell: Wc” , which I found on Hacker News , I thought D could do better. So I wrote a wc in D.

The Program

It consists of one file and has 34 lines and 712 characters.

import std.stdio : writefln, File;
import std.algorithm : map, fold, splitter;
import std.range : walkLength;
import std.typecons : Yes;
import std.uni : byCodePoint;

struct Line {
	size_t chars;
	size_t words;
}

struct Output {
	size_t lines;
	size_t words;
	size_t chars;
}

Output combine(Output a, Line b) pure nothrow {
	return Output(a.lines + 1, a.words + b.words, a.chars + b.chars);
}

Line toLine(char[] l) pure {
	return Line(l.byCodePoint.walkLength, l.splitter.walkLength);
}

void main(string[] args) {
	auto f = File(args[1]);
	Output o = f
		.byLine(Yes.keepTerminator)
		.map!(l => toLine(l))
		.fold!(combine)(Output(0, 0, 0));

	writefln!"%u %u %u %s"(o.lines, o.words, o.chars, args[1]);
}

Sure, it is using Phobos, D’s standard library , but then why wouldn’t it? Phobos is awesome and ships with every D compiler. The program itself does not contain a single if statement. The Haskell wc implementation has several if statements. The D program, apart from the main function, contains three tiny functions. I could have easily put all the functionally in one range chain, but then it probably would have exceeded 80 characters per line. That’s a major code-smell.

The Performance

Is the D wc faster than the coreutils wc ? No, but it took me 15 minutes to write mine (I had to search for walkLength , because I forgot its name).

Memory:

Is the Haskell wc faster? For big files, absolutely, but then it is using threads. For small files, GNU’s coreutils still beats the competition. At this stage my version is very likely IO bound, and it’s fast enough anyway.

I’ll not claim that one language is faster than another. If you spend a chunk of time on optimizing a micro-benchmark, you are likely going to beat the competition. That’s not real life. But I will claim that functional programming in D gives functional programming in Haskell a run for its money.

A Bit About Ranges

A range is an abstraction that you can consume through iteration without consuming the underlying collection (if there is one). Technically, a range can be a struct or a class that adheres to one of a handful of Range interfaces. The most basic form, the InputRange , requires the function

void popFront();

and two members or properties:

T front;
bool empty;

T is the generic type of the elements the range iterates.

In D, ranges are special in a way that other objects are not. When a range is given to a foreach statement, the compiler does a little rewrite.

foreach (e; range) { ... }

is rewritten to

for (auto __r = range; !__r.empty; __r.popFront()) {
    auto e = __r.front;
    ...
}

auto e = infers the type and is equivalent to T e = .

Given this knowledge, building a range that can be used by foreach is easy.

struct Iota {
	int front;
	int end;

	@property bool empty() const {
		return this.front == this.end;
	}

	void popFront() {
		++this.front;
	}
}

unittest {
	import std.stdio;
	foreach(it; Iota(0, 10)) {
		writeln(it);
	}
}

Iota is a very simple range. It functions as a generator, having no underlying collection. It iterates integers from front to end in steps of one. This snippet introduces a little bit of D syntax.

@property bool empty() const {

The @property attribute allows us to use the function empty the same way as a member variable (calling the function without the parenthesis). The trailing const means that we don’t modify any data of the instance we call empty on. The built-in unit test prints the numbers 0 to 10 .

Another small feature is the lack of an explicit constructor. The struct Iota has two member variables of type int . In the foreach statement in the test, we create an Iota instance as if it had a constructor that takes two int s. This isa struct literal. When the D compiler