Copyright (c) 2016 Bart Massey
Herein lie solutions to all of the problems of the 2016
Advent of Code. Advent of Code is
a fantastic exercise, and I thank the author and others
involved profusely for their excellent work. Thanks also to
relsqui for pointing me at this last year.
The solutions are in directories named day01 through
day25. For each solution, I have included commented and
cleaned-up Rust code. There is a README.md in every
problem directory containing descriptions, comments and
usage instructions. I have also included the problem
descriptions (part1.md and part2.md) and my specific
input.txt for posterity.
There is also a libaoc directory containing a library used
by all solutions. It includes a mishmash of AoC-specific
stuff and generally-useful functionality. Please see the
libaoc rustdoc for the details.
There is a talk in rust-aoc.p.md that says some things
about all this.
There are not many tests written for this code. I regard passing both parts of a day's problem as strong validation. But more should be there.
I assume you have Rust running on a fast-ish UNIX box with a bunch of memory (although most everything should also work on other operating systems). A few problems are dependent on common extra packages from Cargo.
The goals of these solutions are to:
-
Provide canonical correct solutions with reasonable runtimes.
-
Illustrate reasonable solution strategies.
-
Illustrate the use of Rust in problem-solving.
I learned a ton of Rust and a little bit of software engineering I should already have known writing these.
There's also some engineering infrastructure in the form of
the template directory, the mkday.sh and
process-aoc.sh shell scripts and the libaoc crate. This
sped up each day's setup considerably. At the beginning of
day 34 I would "sh mkday.sh 34". At the end of the day I
would select and copy the page source of the day 34 AoC page
and then
xclip -selection CLIPBOARD -out | sh ../process-aoc.sh
to get markdown into the problem files for posterity.
These solutions deserve a much more thorough top-level description than I have the energy to write at this point. I will revise this file in the indefinite future.
There is more refactoring that should be done. Specifically:
- The
try_pat()function of days 21 and 22 should be pulled tolibaocand used in other parts as well.
I have run everything through Clippy at this point, and taken much of its advice. I have generally left idiomatic loops alone, as the iterator versions suggested by Clippy seem less clear to me. It was gratifying to see the last three days' solutions pass Clippy without warning: apparently I learned some Rust.
I am a novice Rust programmer. Feedback and pull requests are extremely welcome! Let me know what I should have done, and I'll try to make it right.
This work is licensed under the "MIT License". Please see
the file COPYING in the source distribution of this software
for license terms.
Update 2022-09-02: Revisited this code after getting
automatically warned that some of the ancient crate
dependencies I'd used had security issues. Upgraded and
updated everything, within reason. Added a new library
libcapturesat to paper over a breaking regex crate
change; added an update directory with the shell scripts I
used to quickly go through all 25 days.
I also updated the MD5 digest stuff from rust-crypto,
which is very stale and has security warnings, to md-5,
which is apparently fine.
Hopefully it's better now. I tagged the revision
from 2017 as v2 if you want to see the old stuff.