40 hours a day, now that's some dedicated employees :O

Uiua

Put this one off for a bit and I'll put off part two for even longer because I don't want to deal with any pyramid-shapes of boxes at the moment.
The code for part one feels too long but it works and runs <2s so I'm happy with it for now ^^

Run with example input here

Damn that's a lot time saved. I love how unassuming the addition looks for how great an effect it has

Uiua

Ok, so part one wasn't too hard, and since uiua also takes negative values for accessing arrays, I didn't even have to care about converting my modulus results (though I did later for part two).
I'm a bit conflicted about the way I detected the quadrants the robots are in, or rather the way the creation of the mask-array happens. I basically made a 11x7 field of 0's, then picked out each quadrant and added 1-4 respectively. Uiua's group (⊕) function then takes care of putting all the robots in separate arrays for each quadrant. Simple.

For part two, I didn't even think long before I came here to see other's approaches. The idea to look for the first occurrence where no robots' positions overlapped was my starting point for what follows.

Now on to the more layered approach of how I got my solution.

In my case, there's two occasions of non-overlapping positions before the christmas tree appears.
I had some fun trying to get those frames and kept messing up with going back and forth between 7x11 vs 103x101 fields, often forgetting to adjust the modulus and other parts, so that was great.

In the end, I uploaded my input to the online uiua pad to make visualizing possible frames easier since uiua is able to output media if the arrays match a defined format.

I used this code to find the occurrence of non-overlapping positions (running this locally):

 uiua

    
&rs ∞ &fo "input-14.txt"
⊜(↯2_2⋕regex"-?\\d+")≠@\n.
0 # number of seconds to start at
0_0
⍢(◡(≡(⍜⌵(◿101_103)+°⊟⍜⊡₁×):)◌
  ◿[101_103]≡+[101_103]
  ⊙+₁
| ≠⊙(⧻◴)⧻.)
⊙◌◌
-₁

  

Whenever a new case was found, I put the result into the code in the online pad to check the generated image, and finally got this at the third try:

Welp, got frustrated again with part one because there kept being something wrong with my totally-not-ugly loop and so came here again. I did have to change IsInt (and thus also Cost to account for different handling) for part two though because I kept getting wrong results for my input.
I'm guessing it's because uiua didn't see the difference between rounded and non-rounded number anymore.

Here's the updated, slightly messier version of the two functions that worked out for me in the end :D

 uiua

    
IsInt ← ≍°⊟⍉⍜(⊙(⍉≡↙₂))(/+×)⊙⍉⁅
Cost  ← /+×3_1×⟜IsInt⊸AB


  

_{Could have been done better but I'm lacking the patience for that now}

Nice :D
How's the speed now?

Counting the number of corners was a very useful hint for part 2. I had the most trouble with detecting the double corners, i.e. like in the example where the two B fields touch diagonally:

    
AAAAAA
AAABBA
AAABBA
ABBAAA
ABBAAA
AAAAAA

  

Still, I would've taken a lot longer and probably made really-bad-performance-code without reading this :D

I found multidimensional markers for partition to work really well for finding the fields: Areas ← ⊜□:⇡△.+1⍜♭⊛ It just groups the other array's contents according to adjacent markers, horizontally and vertically. Took me quite a bit to figure out what's actually happening in the example in the documentation ^^'

Uiua

I spent a while thinking about how to best do a flood fill in Uiua when I saw that ⊜ (partition) works beautifully with multidimensional markers: "Groups are formed from markers that are adjacent along any axis.", meaning I just had to convert all letters into numbers and I'd get all indices belonging to a field into an array.
For part 2, I cheated a bit by coming here and reading that you only need to count the edges. To my surprise, the second part is actually a bit faster than part 1. Takes less than 0.2 seconds each though :D

Run with example input here

 uiua

    
$ RRRRIICCFF
$ RRRRIICCCF
$ VVRRRCCFFF
$ VVRCCCJFFF
$ VVVVCJJCFE
$ VVIVCCJJEE
$ VVIIICJJEE
$ MIIIIIJJEE
$ MIIISIJEEE
$ MMMISSJEEE
.
N     ← +[0_¯1 0_1 ¯1_0 1_0]
Areas ← ⊜□:⇡△.+1⍜♭⊛
Peri  ← -/+≡(/+∊N¤)⟜¤⟜(×4⧻)
Sides ← (
  ⊙(-¤)↯:▽⊙0×°⊟.+2⌵⊸-+1⊃⊣⊢⊸⍜⍉≡⍆
  ⧻⊚⊸∊1_3⧈(/+/+)2_2.⍜⊡=₀+1:
  +⊙(×2/+/+⧈(∊[[1_0 0_1][0_1 1_0]])2_2◌)
)
Cost! ← /+≡◇(×^0⟜⧻)

PartOne ← (
  # &rs ∞ &fo "input-12.txt"
  ⊜∘≠@\n.
  Cost!Peri Areas
)

PartTwo ← (
  # &rs ∞ &fo "input-12.txt"
  ⊜∘≠@\n.
  Cost!Sides Areas
)

&p "Day 12:"
&pf "Part 1: "
&p PartOne
&pf "Part 2: "
&p PartTwo

  

Uiua

After finally deciding to put aside Day 9 Part 2 for now, this was really easy actually. The longest was figuring out how many extra dimensions I had to give some arrays and where to remove those again (and how). Then part 2 came along and all I had to do was remove a single character (not removing duplicates when landing on the same field by going different ways from the same starting point). Basically, everything in the parentheses of the Trails! macro was my solution for part 1, just that the ^0 was ◴ (deduplicate). Once that was removed, the solution for part 2 was there as well.

Run with example input here

Note: in order to use the code here for the actual input, you have to replace =₈ with =₅₀ because I was too lazy to make it work with variable array sizes this time.

 uiua

    
$ 89010123
$ 78121874
$ 87430965
$ 96549874
$ 45678903
$ 32019012
$ 01329801
$ 10456732
.
Adj ← ¤[0_¯1 0_1 ¯1_0 1_0]

Trails! ← (
  ⊚=0.
  ⊙¤
  ≡(□¤)
  1
  ⍥(⊙(≡(□^0/⊂≡(+¤)⊙¤°□)⊙Adj
      ≡(□▽¬≡/++⊃=₋₁=₈.°□))
    +1⟜⊸⍚(▽=⊙(:⟜⊡))
  )9
  ⊙◌◌
  ⧻/◇⊂
)

PartOne ← (
  # &rs ∞ &fo "input-10.txt"
  ⊜∵⋕≠@\n.
  Trails!◴
)

PartTwo ← (
  # &rs ∞ &fo "input-10.txt"
  ⊜∵⋕≠@\n.
  Trails!∘
)

&p "Day 10:"
&pf "Part 1: "
&p PartOne
&pf "Part 2: "
&p PartTwo

  

That's also possible, though I've had cases where putting a ? before a function changed the output to what I expected instead of doing something else.
This only happened in the online pad and seems to have been fixed by reloading the tab but I've taken to call any such behavior a bug now :D
Especially in this case because ⋕"1019" should at least not throw a "invalid float literal" error.

Though if you ever find another explanation I'd be happy to read it ^^

(I'm going to 'investigate' some more because I had this error a few times but I don't remember the circumstances and solution anymore)

Great explanation :D
I understand what I called black magic before now. I did wonder what something other than a concatenation would do in that place but didn't consider that it might be just that because it looked so complicated.

I wasn't able to get it running with ⋕$"__" either, though I'd assume it's a bug caused by how the numbers get passed around or something. The day uiua has a stable release will be glorious

Uiua

Adapting the part one solution for part two took me longer than part one did today, but I didn't want to change much anymore.

I even got scolded by the interpreter to split the evaluating line onto multiple ones because it got too long.
Can't say it's pretty but it does it's job ^^'

Run with example input here

 uiua

    
PartOne ← (
  &rs ∞ &fo "input-8.txt"
  ⟜(▽¬∈".\n".◴)
  ⊜∘≠@\n.
  :¤⟜(:¤-1△)
  ≡(□⊚⌕)
  ◴/◇⊂⍚(≡(-:⟜-°⊟)⧅≠2)
  ⧻▽¬:⊙(/+⍉+)⟜⊓><,0
)

PartTwo ← (
  &rs ∞ &fo "input-8.txt"
  ⟜(▽¬∈".\n".◴⟜¤
    ▽:⟜≡(>1⧻⊚⌕)
  )
  ⊜∘≠@\n.
  :¤⟜(:¤-1△)
  ≡(□⊚⌕)
  ⊸⍚(
    ⧅≠2⊙¤
    ≡(:¤⟜-°⊟
      ⍢(⊙⊂⟜-⊙⊸⊢
      | ⋅(=0/++⊓><,0⊢))
      □⊙◌◌
    )
  )
  ◴/◇⊂/◇⊂
  ⧻▽¬:⊙(/+⍉+)⟜⊓><,0
)

&p "Day 8:"
&pf "Part 1: "
&p PartOne
&pf "Part 2: "
&p PartTwo

  

A solution in malbolge would be amazing and also kinda terrifying

You can actually use (singular) emojis as variable names. According to the documentation, they won't be be used as glyphs so you don't even have to worry about breaking changes of that kind :D

Hell yeah!

┻━┻︵ °□°)/ ︵ ┻━┻

No worries, it does seem a lot less difficult in hindsight now, my mind just blanked at what I expected to be a lot more code :))

That performance improvement is amazing, I'll definitely take a look at how that works in detail later. Just gotta recover from the mental stretch gymnastics trying to remember the state of the stack at different code positions

Glad to hear that my attempts at de-spaghettifying worked to some degree at least :D

Uiua, forgot to put that in the post

Thanks to your solution I learned more about how to use reduce :D

My solution did work for the example input but not for the actual one. When I went here and saw this tiny code block and you saying

This turned out to be reasonably easy

I was quite taken aback. And it's so much better performance-wise too :D (well, until part 2 comes along in my case. Whatever this black magic is you used there is too high for my fried brain atm)