Was contemplating doing this year's in Zig (just because) or Kotlin (due to some work stuff), but this december is CRAZYpants (at least in the beginning), so will fall back to comfy languages.
Welp, I was going to do Swift, but yay:
% swift sh
error: unable to invoke subcommand: swift-sh (No such file or directory)
So, Objective-C it is until I can get swift fixed on this machine. sigh
Update: looks like past-me in 2021 used https://nshipster.com/swift-sh/, for easier running of swift files as scripts.
==================================================
Build a graph, run through instructions. Seems straightforward. For parsing, tried regexs, but swift kept complaining about it, so going to do it stupid-style
==================================================
Loks like the complexity is detecting poker hands. There's only 1000 lines in the prod databases, so probably can just brute force it like before. Kind of nice mid-week not having to think too hard.
Score and accumulate ranks and bids. Hardest part was reading the requirements. PEBCAK
jokers! Not in the mood, so waiting on this one til later.
==================================================
I think it's just "do the same math over a range". Since it feels bell-curvey, could probably optimize by approaching from the ends.
Nah, brute force worked fine
just brute force over :all-the-things:? Some of the underdogs talking about quadratic formula. Brute force took all of a minute.
==================================================
Get a feeling that "just understanding the problem is enough to solve it", with a non-exhaustive data structure, since the prod data uses BIG numbers
yeah. just a little tedious cascading through all the maps
arg, these take forever to run - prod data has a redonk number of things to chew throw
stride 114805397 stride 164506173 stride 166087295 stride 340945548 stride 483360452 stride 252097746 stride 279314434 stride 145640027 stride 149488635 stride 236791935
the solution I got was too high 17729183. I probably have an obiwan in the map code that's not getting exercised by the inidividual seed numbers.
=================================================
looks like a simple "stuff things into a lookup table. look other things up and calculate a value". I bet the card number value will be involved in the twist.
Pretty straightforward
Took longer to understand the problem than to do it :-)
==================================================
Looks like adjacency of ranges is the crux. yep!
Kind of over-designed pulling numbers and symbols out into their own structures, and making stripes of arrays for each line. Hopefully the twist won't mess me up.
The over-engineering paid off. essentially swapping the loop that finds adjancy overlaps, and done.
https://adventofcode.com/2023/day/2
Looks like :alot: of spltting and comparing to a fixed set of nubmers.
yeah, pretty plug and chug
Looks like find the max of each color. yep! Much easier than yesterday.
https://adventofcode.com/2023/day/1
Looks like a "scan string from ends, make a number"
1abc2 == 12
pqr3stu8vwx == 38
a1b2c3d4e5f == 25
treb7uchet == 7
Let's take a moment and marvel at the examples, that answered all the questions I had.
And yay, got it right.
lol. that's hilarious. Need to look up character strings as well.
two1nine = 29
eightwothree = 83
abcone2threexyz = 13
xtwone3four = 24
4nineeightseven2 = 42
zoneight234 = 14
7pqrstsixteen = 76
(kind of tedious...)
hrm. eyeballing, seems to work fine. I am parsing eighthree as 83 and sevenine as 79, but saying I'm too low.
Heh. Ran it through a random python script from reddit, and it got a lower value (by six). Also wrong.
Found a "thought there'd be a pear, but nothing there"
original: 55336 with a stupid memory fix: 30093 python script: 55330 found a C++ version 55343
ok, the problem was not getting repeated words