Advent of Code 2025 Discussions

General
31

Using in-combinations makes things a bit cleaner (and faster, I suspect), which would've been useful on day 8 as well, but I had forgotten about it.

Spoilers 
; day 9
; part 1
(for/fold ([best -inf.0])
          ([pair (in-combinations red-tiles 2)])
  (max best (apply area pair)))
; part 2
(for/fold ([best -inf.0])
          ([pair (in-combinations red-tiles 2)]
           #:do [(define r (apply bounds pair))]
           #:when (andmap in-polygon? (rectangle-frame r))
           #:unless (bisected? (rectangle-around? r)))
  (max best (rectangle-area r)))

; day 8
(define pairwise-distances
  (sort
   #:key car
   (for/list ([pair (in-combinations junctions 2)])
     (cons (apply distance pair) pair))
   <))

As an aside, I started using complex numbers because I thought they would be useful when calculating areas, but I couldn't quite get the calculations to work, so I ended up deconstructing them anyhow. They were useful for the interpolation, though, which was fortunate.

Edit: I realized the area comparison could be reordered to speed it up a bit.

Spoilers 
(time
 (for/fold ([best -inf.0])
           ([pair (in-combinations red-tiles 2)]
            #:do [(define r (apply bounds pair))]
            #:when (< best (rectangle-area r))
            #:when (andmap in-hull? (rectangle-corners r))
            #:unless (hull-bisects? (rectangle-contains? r)))
   (rectangle-area r)))

;=> cpu time: 1343 real time: 1383 gc time: 468

Although this doesn't show the optimization for interpolated points, which I also realized only really needed to be calculated once.

1 Like