Giving up for today, still working on part 1. My code handles the examples correctly, but is terribly slow on the real input, a combination of doing too much work with way too much memory. More tomorrow, though I might do day 16 first

Same here, although my almost working solution is not that slow even on full input.

Then I'll be interested to see yours once it's finished! I really can't think of an elegant way to make it faster.

@karelpeeters Day15 - Part1 solution is here which runs in <1sec. Still getting incorrect answer for Part 2 & no time to resolve. https://gitlab.com/abowes/adventofcode2018/blob/master/src/main/kotlin/jfactory/day15/BeverageBandits.kt

Nice clean code! Two things I noticed that might help with the second part. First, it seems to me your stop condition for part 1 looks if there are still enemies alive of the first sprite in the list. I think this only works if the first sprite is alive? Second, there is some weird logic about how to count full rounds, which I don't think you implemented. Essentially, if a army wins halfway through a round, that round shouldn't count towards the total.

That's this part, no? https://gitlab.com/abowes/adventofcode2018/blob/master/src/main/kotlin/jfactory/day15/BeverageBandits.kt#L111

Yes, and the same call a few lines above on line 103

So doesn't he only count the current round if there are still enemies alive?

You're absolutely right. And, my other point is also invalid, because I now see on line 112 that dead sprites get removed. So, wrong on both counts, I was!