I'm having an intense issue and tried looking thro...

# advent-of-code

[Tolly Kulczycki]

I'm having an intense issue and tried looking through the no-spoilers thread but no dice, need help, code in thread

[Tolly Kulczycki]

Untitled.cpp

Untitled.cpp

[Tolly Kulczycki]

my approach is to go forward with ranges of seeds

[Tolly Kulczycki]

I imagine it's an off-by-one error... somewhere in there

[Dan Fingal-Surma]

seedRanges.addAll(ranges.second) does not seem correct, that puts the transformed range back into the queue to be processed from the beginning as if it were a seed range

[Dan Fingal-Surma]

you want to proceed in stages, take all the seed input ranges and produce the soil ranges, then take the soil ranges and produce the fertilizer ranges, etc.

[Dan Fingal-Surma]

Looking at transformRange I don’t see where it projects it into the destination space

[Dan Fingal-Surma]

I had this error at one point as well

[Dan Fingal-Surma]

you still need your transform function from part 1, currSeed = option[0] + (currSeed - option[1])

[Dan Fingal-Surma]

the idea to move forward with ranges is the key insight which you already have

[Tolly Kulczycki]

so for me mapping is this:

60 56 37
56 93 4

and an option is this:

60 56 37

[Tolly Kulczycki]

and so transform range takes a range and produces a list of ranges that

[Tolly Kulczycki]

are either outside below, outside above, or modified

[Tolly Kulczycki]

and the ones outside get back onto the stack

[Tolly Kulczycki]

while if something has been modified then it breaks

[Tolly Kulczycki]

and moves onto the next stage

[Dan Fingal-Surma]

how much spoilers do you want in terms of the algorithm

[Tolly Kulczycki]

I can tolerate some, already looked at some of the answers

[Dan Fingal-Surma]

so at phase N, for each input range, you want to produce the set of output ranges. then feed all produced output ranges into phase N + 1.

If you put it back into the deque, it will start processing that range from phase 0, which is not correct. i.e. applying a seed-to-soil transform on a light range.

[Tolly Kulczycki]

OH

[Tolly Kulczycki]

That's the problem

[Tolly Kulczycki]

I see now

[Tolly Kulczycki]

okay

[Dan Fingal-Surma]

producing the set of output ranges for a given input range has 2 steps, at least in my impl: determine the interior split points based on the mapping ranges, which generates a set of (contiguous) ranges, then project those ranges into the destination space

[Tolly Kulczycki]

That makes so much sense, thank you Dan!

[Tolly Kulczycki]

Okay, managed to figure it out (albeit with your knowledge, another tutorial and some reading) but I did it!