release-1.10: bump SparseArrays to latest v1.10#56543
release-1.10: bump SparseArrays to latest v1.10#56543KristofferC wants to merge 3 commits intorelease-1.10from
Conversation
|
@gbaraldi, @ViralBShah, @wsmoses, seems like this changed what error is thrown (SparseArrays does type piracy). |
|
I have no idea why that would change, but looking at the test, if anything that error feels more consistent? @testset "splatting into hvcat" begin
t = (1, 2)
@test [t...; 3 4] == [1 2; 3 4]
@test [0 t...; t... 0] == [0 1 2; 1 2 0]
@test_throws ArgumentError [t...; 3 4 5]
@test Int[t...; 3 4] == [1 2; 3 4]
@test Int[0 t...; t... 0] == [0 1 2; 1 2 0]
@test_throws DimensionMismatch Int[t...; 3 4 5]
end |
|
lmao wat |
|
It depends on if SparseArrays have been loaded or not into the worker because SparseArrays does type piracy on So the error thrown has to be consistent between Base and SparseArrays. |
|
okay one error path is: The other error is To me this is a bug, why should typed_hvcat throw a different error from hvcat for the same issue? |
|
okay @KristofferC I think JuliaSparse/SparseArrays.jl#588 ought to do it but I'm a bit unclear on how to set up the end to end testing here |
|
Okay I think that error type we changed in this PR now should be restored to the base Julia version. id update it but I don’t have permissions |
|
Added to #56653. |
No description provided.