dijkstra_reach

[danielhuang]

There's still some potential bugs; will need to test more.

[codspeed-hq]

CodSpeed Performance Report

Merging #462 will not alter performance

_{Comparing danielhuang:dijkstra-reach (1406608) with main (3372900)}

Summary

✅ 32 untouched benchmarks

[samueltardieu]

Thanks for this work-in-progress. What problem is it trying to solve that would not be solved using bfs_reach() or dfs_reach()? Do you have any significant problem where this would make a difference?

[danielhuang]

dijkstra_reach is used in my optimized 2022 day 19 solution, which finishes in 300ms (there's still some improvements to be made, since hashing seems to be the bottleneck here).

[samueltardieu]

300ms is a great time for this problem!

[samueltardieu]

I wonder which code could be factored out to avoid a repetition with other dijkstra…() methods

[samueltardieu]

@danielhuang Did you try to iron out remaining bugs (you said there could be some of them left)?

[danielhuang]

Most of the bugs should be gone now; though I haven't done any fuzzing—only pre-written tests.

[samueltardieu]

Most of the bugs should be gone now; though I haven't done any fuzzing—only pre-written tests.

Ok, can you squash your commits into one with subject "feat: add dijkstra_reach()" and force push? I'll then add some more tests, documentation, and merge it. Thanks!

[samueltardieu]

@danielhuang Any news? It would be great to integrate your patch before the AoC 2023.

[samueltardieu]

I've squashed the commits and added some documentation

[samueltardieu]

Integrated in the just released 4.4 version. Thanks!