Advent of Code 2024 in Rust

I'm trying the advent of code in Rust again this year to learn more Rust. I'm not trying to be fast and place on the leaderboards (which also require working at midnight, which I'm mostly not interested in doing), I'm trying to be somewhat elegant and learn new things in Rust. The documentation is live here.

The focus of these solutions is a balance of speed and readability.

I highly recommend loading this up in a good editor, like Visual Studio Code or VIM with the ALE plugin. It will add type information to all inferred types, autocomplete, show documentation, etc.

Feel free to check out my 2023 repo at https://github.com/henryiii/aoc2023.

Formatting and linting

Use:

cargo fmt --all && cargo clippy-all

If you want to auto-fix anything, you can:

cargo clippy --fix --allow-dirty --allow-staged

I also looked for removable features using unused-features, both to speed up compilation and it helped removed a small dependence on unicode in regex, though it doesn't handle workspaces, sadly.

Tests

Use:

cargo test --package year_2024

Useful flags include -- --nocapture and -- day_<NUMBER> for just one set of tests.

If you have cargo-nextest (say, from brew install cargo-nextest), then cargo nextest run also works.

New days

To create a new day, use:

cargo newday 01

This will make three files for you, the src/bin file, the samples file, and the data file.

Downloading input

Download the input files to data/<number>.txt. For example, data/01.txt.

To download the input from the CLI, use:

cargo getdata 2024 01 --session $AOC_TOKEN > data/2024/01.txt

The AOC token is in your cookies, and you can get it from the browser. You can set the environment variable AOC_TOKEN to avoid typing it each time.

The newday command above also accepts --session/AOC_TOKEN, which will cause it to download the input for you if it's available.

See https://github.com/gobanos/cargo-aoc which inspired this feature and has more info about getting your token. In Safari, it's in the devtools inspection pane.

Running

Use:

cargo 2024 01

A few days accept --vis to add a visual.

Docs

You can build with:

cargo docs --no-deps --workspace

Profiling

I've used samply to profile. After installing (cargo install --locked samply), you can use samply record target/bin/XX and open the page in Chrome or Firefox. I've enabled debug info in release mode for this. You can rebuild Rust's stdlib with debuginfo for more fine grained detail, but I haven't needed that. I have replaced par_iter with iter when doing this, otherwise Rayon hides what's happening. I could use a feature to automate this, but I think keeping usage simple is fine.

Debugging

To use lldb, just do the following:

$ cargo build --package year_2024 --bin XX
$ lldb target/debug/XX
> breakpoint set -f crates/year_2024/src/XX.rs -l YY
> r

It should work like any other executable, mostly. The visual debugger in VSCode might work, though I had a hard time telling it to run the correct executable. Note that LLDB only supports the subset of Rust and C for conditional breakpoints, so you might have to manually add code and use line based breakpoints instead.

Notes

Also see Blessed.rs, a curated list of good Rust libraries.

Final timings

On my MacBook (Intel), these are my timings (total program runtime 452ms, was 1223ms originally):

Day 01
Solution A: 1197984    (0.154ms)
Solution B: 23387399    (0.197ms)
Time taken: 0.448ms
Day 02
Solution A: 486    (0.511ms)
Solution B: 540    (0.578ms)
Time taken: 1.150ms
Day 03
Solution A: 165225049    (0.589ms)
Solution B: 108830766    (0.641ms)
Time taken: 1.293ms
Day 04
Solution A: 2464    (0.980ms)
Solution B: 1982    (1.276ms)
Time taken: 2.329ms
Day 05
Solution A: 5208    (1.286ms)
Solution B: 6732    (1.330ms)
Time taken: 2.739ms
Day 06
Solution A: 5312    (0.183ms)
Solution B: 1748    (23.254ms)
Time taken: 23.529ms
Day 07
Solution A: 12940396350192    (1.027ms)
Solution B: 106016739583593    (6.177ms)
Time taken: 7.346ms
Day 08
Solution A: 228    (0.112ms)
Solution B: 766    (0.126ms)
Time taken: 0.321ms
Day 09
Solution A: 6471961544878    (6.063ms)
Solution B: 6511178035564    (23.227ms)
Time taken: 29.362ms
Day 10
Solution A: 794    (1.510ms)
Solution B: 1706    (1.366ms)
Time taken: 2.960ms
Day 11
Solution A: 194782    (0.211ms)
Solution B: 233007586663131    (7.184ms)
Time taken: 7.451ms
Day 12
Solution A: 1477762    (14.494ms)
Solution B: 923480    (18.124ms)
Time taken: 32.714ms
Day 13
Solution A: 27157    (0.609ms)
Solution B: 104015411578548    (0.258ms)
Time taken: 0.993ms
Day 14
Solution A: 228457125    (0.215ms)
Solution B: 6493    (26.046ms)
Time taken: 26.334ms
Day 15
Solution A: 1457740    (1.523ms)
Solution B: 1467145    (8.126ms)
Time taken: 9.767ms
Day 16
Solution A: 72400    (15.611ms)
Solution B: 435    (23.902ms)
Time taken: 39.624ms
Day 17
Solution A: 7,4,2,5,1,4,6,0,4    (0.028ms)
Solution B: 164278764924605    (0.106ms)
Time taken: 0.220ms
Day 18
Solution A: 320    (4.971ms)
Solution B: 34,40    (21.417ms)
Time taken: 26.468ms
Day 19
Solution A: 300    (25.684ms)
Solution B: 624802218898092    (64.211ms)
Time taken: 90.012ms
Day 20
Solution A: 1524    (15.323ms)
Solution B: 1033746    (15.942ms)
Time taken: 31.360ms
Day 21
Solution A: 137870    (0.069ms)
Solution B: 170279148659464    (0.175ms)
Time taken: 0.319ms
Day 22
Solution A: 18694566361    (9.192ms)
Solution B: 2100    (35.625ms)
Time taken: 44.902ms
Day 23
Solution A: 1512    (34.062ms)
Solution B: ac,ed,fh,kd,lf,mb,om,pe,qt,uo,uy,vr,wg    (18.589ms)
Time taken: 52.754ms
Day 24
Solution A: 57270694330992    (3.152ms)
Solution B: gwh,jct,rcb,wbw,wgb,z09,z21,z39    (0.699ms)
Time taken: 3.916ms
Day 25
Solution A: 3287    (4.662ms)
Solution B: 0    (0.000ms)
Time taken: 4.702ms

Wasm experiments

I'm trying Wasm. I installed wasm-pack. I also installed the wasm32-unknown-unknown target with rustup target add wasm32-unknown-unknown. I then ran wasm-pack build crates/wasm_2024 --target no-modules.