Kristoffer Carlsson

The `download_artifacts` is taken care of by https://github.com/JuliaLang/Pkg.jl/pull/4117.

Random thought, would it make sense to implement the https://github.com/JuliaCollections/AbstractTrees.jl interface for the trees in this package?

As a check of the functionality here it would be nice to reimplement https://github.com/KristofferC/NearestNeighbors.jl/blob/master/examples/balltree_illustration.ipynb using these official traverse functions. Doesn't strictly have to be done here but it would serve...

Maybe you are aware but I want to point out https://gist.github.com/kateinoigakukun/b0bc920e587851bfffa98b9e279175f2 so you can keep the blame functionality working without everything pointing to this PR modifying it.

I'm saying that based on that CI here was never tested with the Base changes to lazy jlls. As far as I see, the CI status when the lazy jll...

That's not the issue. This is the issue https://buildkite.com/julialang/linearalgebra-dot-jl/builds/377#0195e5fe-f561-4369-b6ff-6fcdb72a05d4 ``` LinearAlgebra/triangular (16) \| started at 2025-03-30T11:02:56.257 --   | From worker 9:   | From worker 9: [537] signal 11 (1): Segmentation...

Not straight forward. ForwardDiff differentiates w.r.t numbers and abstract vectors. You might be able to hack something together with generated functions.

Yep, seems like a straight up bug. Happy for a PR.

I do like that you don't need to duplicate the assembler in the scratch value as done now. Imo I think we should go with this assuming the overhead can...

I don't think this really has to be on the milestone