I enjoyed this one a ton! I’m pretty happy with how my code turned out: https://gist.github.com/QuinnWilton/35d82ae4aab4fb39e2ac622ab54fc460
I liked building the interpreter, but hopefully it was the last time we had to use it. 
@qhwa Slick solution. 
1 Like
And I hope for more of them, hoping to be able to maintain compatibility from day 2 until day 25 (or whenever the last update is done)
Today I used a couple of techs learned from your last video, such as the combination of rem and div to get digits.
And interestingly, I found some common mistakes we both made, such as wrongly counting the pointer. 
It was a long puzzle and bugs happened. I am still stuck in Y2015 Day22, passing all the tests yet not able to figure out the solution. 
Helpfully this time a lot of test codes were present, making things a lot easier.
So, thanks for sharing, again! Looking forward to more of it.
1 Like
Struggled a little with part2 of Day 5. Wrote a disassembler to see what was going on. Hopefully it will be useful for days ahead.
Disassembed Listing: input_05.txt
Program Size: 678
Entry Point: 0
Input List: 5
----------
0: input 5
2: add 5, 1100, 6
6: jmpt 1, 238
238: jmpt 0, 99999
241: jmpt 227, 247
247: jmpt 0, 99999
250: jmpt 3, 256
256: jmpf 227, 99999
259: jmpf 0, 265
265: jmpf 3, 99999
268: jmpf 0, 274
274: jmpt 1, 280
280: add 5, 5, 225
284: add 294, 0, 0
288: jmpt 1, 294
294: jmpf 0, 300
300: add 10, 10, 225
304: add 314, 0, 0
308: jmpf 0, 314
314: cmpeq 226, 677, 224 ; 0 written
318: mul 2, 0, 223
322: jmpf 0, 329
329: cmpless 677, 226, 224 ; 0 written
333: mul 0, 2, 223
337: jmpf 0, 344
344: cmpeq 677, 226, 224 ; 0 written
348: mul 2, 0, 223
352: jmpf 0, 359
359: cmpless 226, 677, 224 ; 1 written
363: mul 2, 0, 223
367: jmpt 1, 374
374: cmpeq 677, 226, 224 ; 0 written
378: mul 0, 2, 223
382: jmpf 0, 389
389: cmpless 677, 677, 224 ; 0 written
393: mul 2, 0, 223
397: jmpf 0, 404
404: cmpless 226, 677, 224 ; 1 written
408: mul 0, 2, 223
412: jmpf 1, 419
415: add 0, 1, 223
419: cmpeq 677, 226, 224 ; 0 written
423: mul 2, 1, 223
427: jmpt 0, 434
430: add 2, 1, 223
434: cmpeq 226, 226, 224 ; 1 written
438: mul 3, 2, 223
442: jmpf 1, 449
445: add 6, 1, 223
449: cmpless 677, 226, 224 ; 0 written
453: mul 7, 2, 223
457: jmpf 0, 464
464: cmpless 226, 677, 224 ; 1 written
468: mul 2, 14, 223
472: jmpt 1, 479
479: cmpless 677, 226, 224 ; 0 written
483: mul 28, 2, 223
487: jmpt 0, 494
490: add 56, 1, 223
494: cmpeq 226, 677, 224 ; 0 written
498: mul 57, 2, 223
502: jmpt 0, 509
505: add 114, 1, 223
509: cmpless 677, 677, 224 ; 0 written
513: mul 115, 2, 223
517: jmpf 0, 524
524: cmpless 677, 677, 224 ; 0 written
528: mul 2, 230, 223
532: jmpt 0, 539
535: add 1, 460, 223
539: cmpless 677, 226, 224 ; 0 written
543: mul 2, 461, 223
547: jmpt 0, 554
550: add 922, 1, 223
554: cmpeq 677, 677, 224 ; 1 written
558: mul 923, 2, 223
562: jmpf 1, 569
565: add 1846, 1, 223
569: cmpeq 226, 677, 224 ; 0 written
573: mul 1847, 2, 223
577: jmpf 0, 584
584: cmpeq 226, 226, 224 ; 1 written
588: mul 3694, 2, 223
592: jmpf 1, 599
595: add 7388, 1, 223
599: cmpeq 677, 226, 224 ; 0 written
603: mul 2, 7389, 223
607: jmpf 0, 614
614: cmpless 226, 226, 224 ; 0 written
618: mul 14778, 2, 223
622: jmpf 0, 629
629: cmpeq 226, 677, 224 ; 0 written
633: mul 2, 29556, 223
637: jmpt 0, 644
640: add 1, 59112, 223
644: cmpless 226, 677, 224 ; 1 written
648: mul 59113, 2, 223
652: jmpt 1, 659
659: cmpeq 226, 226, 224 ; 1 written
663: mul 2, 118226, 223
667: jmpf 1, 674
670: add 236452, 1, 223
674: output 236453
676: stop
----------
Outputs Written: 236453
5 Likes
I used Elixir for some problems but not day 2 and 5 as mutability seems to be very practical here.
I did not felt like creating a new memory structure for each write. I was tempted to use ETS though but it seemed overkill.
Tonight I took some time and tried to screencast myself while refactoring my intcode machine.
It is currently uploading to youtube and should be available soon:
Be aware though that it is 90 minutes of me beeing desoriented stroling through my own code talking with myself through cheap bluetooth headphones with an even cheaper microphone, because it was the only thing that I got to work which didn’t make more background noise than my voice in the foreground…
3 Likes
Wow, I really love your solution @qhwa. I was so focused on decomposing everything into small functions, but I think that you’ve demonstrated how that’s really an over-complication of this problem. I think that the amount of coupling and inter-dependency between the “components” are high enough here that defining functions locally with state in context is actually more elegant and simple than decomposition.
For comparison, this was my solution (a total refactor of my messier earlier attempt): https://github.com/montanonic/advent2019_elxir/blob/master/lib/day5.ex
I drew very heavily from Sasa’s solution after debugging older code for days. I hadn’t thought to store input and output data in a record along with the program instructions, and learned from his solution that the design benefits heavily from passing around a larger state record here. But I think again, what this really points at, is how interdependent the parts are on the overall state.
1 Like
