Advent of Coco 2024

A year after the first attempt to do Advent of Code in Coco programming language, here come the new examples. Now with Coco compiler v0.6.0 that supports packages, and has integrated execution of PISA code, programming and running the code got much easier. And using Advent of Code to find compiler bugs is still very much the main aim of this exercise.

The code can be run with

coco test --debug_pisa

as the correct results are parts of the test. So, the numbers are in .coco files, but the fun is getting to them by running the code.

Day1

The results are already in the code, find the lines in d01.coco

// > distance: ...
// > similarity: ...

To run the code and see if the results are OK, do this:

cd AoCoco2024/01
coco test --debug_pisa

Advent of Coco 2023

Here's an attempt to do https://adventofcode.com in Coco programming language (https://cocolang.dev).

It's unlikely this will be doable with anything complex, but having at least the first day completed, it's a proof Coco can work as a general programming language.

And of course, an excellent way to find Coco compiler bugs!

Day1

If you're only interested in two numbers, install Coco 0.3.5 and run:

cd 01
coco compile
logiclab run input1.lab | grep calibration
logiclab run input2.lab | grep calibration

If you're interested in more details, read on!

Running

Install Coco compiler, release 0.3.5 from https://cocolang.dev/releases

Results for Day1 are well, in the folder 01. You need to initialize logiclab before first run (if there's not ./logiclab folder).

cd 01
logiclab init

To invoke (=run) any function (endpoint) in Logiclab, a sender has to be designated. This is done only once. First start the logiclab with logiclab start and then issue two commands:

register TestUser
set designated.sender TestUser

Test

As Coco enables testing out-of-the box, the code includes tests from the exercise in the source code, so we can verify correctness. Default base fuel in logiclab is 10000, which is far too little for loops, so even the simplest test need higher fuel (1,000,000 is a good start).

coco labtest --fuel 1000000

Exercise

First, we need to compile the code, coco.nut file determines the format and output file (d01.yaml)

coco compile

The real exercises require long inputs, as Coco can't read files, we have to provide the input as a long function argument. Luckily, logiclab can handle it. We need to set an even larger fuel limit as we did for the previous test, read and compile the manifest and invoke Calibrate or Calibrate2 endpoints. All commands are written in input1.lab and input2.lab files, so we can just do

logiclab run input1.lab | grep calibration
logiclab run input2.lab | grep calibration

Day2

Both answers are in the same script, so:

cd 02
coco compile
logiclab run input1.lab | grep sum