Advent of Code, 2022
Advent of Code, 2022
These are my advent of code solutions. In the same way as last year and the year before that, I do every day in a different language.
- Must do all days in different languages, except those days which continue from other days' solutions (e.g. IntCode computer, 2019)
- If I do a day in multiple different languages, I can "pick" the language for that day, as long as the resulting picks adhere to rule 1.
- If I wish to use a language that I have already used, I can go back and redo the day that language was previously used for (in a different language), thus freeing it up for the new day.
- Languages cannot be too similar, keeping in spirit of the rules (e.g. bash/zsh, typescript/javascript, haskell/purescript [controversial, I know])
- I may write preprocessors to transform the input into something palatable to the language, but generally this is discouraged.
Languages and solutions so far can be got by running ./summarize.
|1|2|3|4|5|6|7|
awk |█|█|█| | | | |
brainfuck | | | | | |█| |
dc |█| | | | | | |
forth | |█| | | | | |
haskell | |█| | | | | |
perl | | | |█| | | |
sed | | |█| | | | |
sh |█|█| | | | |█|
vim | | | | |█| | |
In my summaries, I will go over the solutions I find more interesting, rather than detailing each language for each day.
Day 1 was in Awk, which is a pretty decent language for this sort of thing - running totals and matching on empty lines is a breeze, letting me write a solution with no overhead.
Forth is a really fun fit. We can create 6 functions, A B C and X Y Z, and then evaluate the input as code!
Writing a solution in GNU Sed is tough, but still fun - one can think of it as an L-system with a few bells and whistles. Since it just rewrites strings, we have to reimplement addition via carries, and splitting a string into two pieces is actually the most complicated part.
Perl has excellent matching utilities, and with the -n flag and END blocks, it can behave like a sophisticated form of Awk. Perl is probably the first "real" language I've used as my primary for a day.
No vimscript, just vim macros. The core of this solution is the * key - using
it, we can implement branching, which can be coupled with macros to make a
fairly turing complete system. I learned a lot about how to break out of
macros, and wrote some truly ugly, unreadable line noise to boot. All in all a
good day.
An ideal day for brainfuck: a growing stack and a fixed-size state within that stack. Still extremely slow, of course.
Why not simulate a shell in a shell? I feel a bit icky about the amount of
logic in the awk component for part 2, but a trick using bc would be no
better.