Exponentiation of Diagonal{*Range} fails

[dkarrasch]

julia> using LinearAlgebra; D = Diagonal(1:3)
3×3 Diagonal{Int64, UnitRange{Int64}}:
 1  ⋅  ⋅
 ⋅  2  ⋅
 ⋅  ⋅  3

julia> D^2
ERROR: MethodError: no method matching Vector{Int64}(::Diagonal{Int64, UnitRange{Int64}})
Closest candidates are:
  Array{T, N}(::AbstractArray{S, N}) where {T, N, S} at array.jl:540
  Vector{T}() where T at boot.jl:467
  Vector{T}(::Core.Compiler.AbstractRange{T}) where T at range.jl:1042
  ...
Stacktrace:
 [1] convert(#unused#::Type{Vector{Int64}}, a::Diagonal{Int64, UnitRange{Int64}})
   @ Base ./array.jl:532
 [2] Diagonal{Int64, Vector{Int64}}(diag::Diagonal{Int64, UnitRange{Int64}})
   @ LinearAlgebra /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.6/LinearAlgebra/src/diagonal.jl:10
 [3] convert(T::Type{Diagonal{Int64, Vector{Int64}}}, m::Diagonal{Int64, UnitRange{Int64}})
   @ LinearAlgebra /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.6/LinearAlgebra/src/special.jl:53
 [4] to_power_type(x::Diagonal{Int64, UnitRange{Int64}})
   @ Base ./intfuncs.jl:240
 [5] power_by_squaring(x_::Diagonal{Int64, UnitRange{Int64}}, p::Int64)
   @ Base ./intfuncs.jl:255
 [6] ^(A::Diagonal{Int64, UnitRange{Int64}}, p::Int64)
   @ LinearAlgebra /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.6/LinearAlgebra/src/dense.jl:447
 [7] macro expansion
   @ ./none:0 [inlined]
 [8] literal_pow(f::typeof(^), x::Diagonal{Int64, UnitRange{Int64}}, #unused#::Val{2})
   @ Base ./none:0
 [9] top-level scope
   @ REPL[41]:1

It fails because the Diagonal{T,V} constructor is being called with a diagonal matrix as argument (item [2] in the stacktrace above), whereas the constructor here expects a vector.

Originally posted by @daanhb in #40831 (comment)

[dkarrasch]

Should be easily fixable by

(^)(D::Diagonal, a::Number) = Diagonal(D.diag .^ a)

[ArunS-tack]

Should be easily fixable by

(^)(D::Diagonal, a::Number) = Diagonal(D.diag .^ a)

In which file should this be added/edited?

[andreasnoack]

It should be added to stdlib/LinearAlgebra/src/diagonal.jl søn. 23. maj 2021 kl. 08.58 skrev Arun sanganal ***@***.***>:

…

Should be easily fixable by (^)(D::Diagonal, a::Number) = Diagonal(D.diag .^ a) In which file should this be added/edited? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#40886 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAD3JKMRYXQOSCMP2YBC4DTTPCRPXANCNFSM45GRL67A> .

[mcabbott]

This fails on 1.6, but seems to work on master:

julia> D = Diagonal(1:3);

julia> D ^ 2
3×3 Diagonal{Int64, Vector{Int64}}:
 1  ⋅  ⋅
 ⋅  4  ⋅
 ⋅  ⋅  9

julia> let p = 3; D ^ p; end
3×3 Diagonal{Int64, Vector{Int64}}:
 1  ⋅   ⋅
 ⋅  8   ⋅
 ⋅  ⋅  27

[ArunS-tack]

This fails on 1.6, but seems to work on master:

julia> D = Diagonal(1:3);

julia> D ^ 2
3×3 Diagonal{Int64, Vector{Int64}}:
 1  ⋅  ⋅
 ⋅  4  ⋅
 ⋅  ⋅  9

julia> let p = 3; D ^ p; end
3×3 Diagonal{Int64, Vector{Int64}}:
 1  ⋅   ⋅
 ⋅  8   ⋅
 ⋅  ⋅  27

How do i test it? :/ I m new to all this. I am still trying to figure out how to Test the changes before making a PR.

[mcabbott]

You can just paste exactly what you quoted into the REPL. You can download a the "nightly" version of Julia from the main download page, and keep this alongside 1.6 on your computer, to try things out.

Actual tests for a PR, you should find the right file in test/ & add some lines (or a new testset) there. But this too you can try out locally, by just using Test and pasting your new tests (and any nearby ones you are worried about).

[ArunS-tack]

You can just paste exactly what you quoted into the REPL. You can download a the "nightly" version of Julia from the main download page, and keep this alongside 1.6 on your computer, to try things out.

Actual tests for a PR, you should find the right file in test/ & add some lines (or a new testset) there. But this too you can try out locally, by just using Test and pasting your new tests (and any nearby ones you are worried about).

In this case, paste the whole diagonal.jl code into repl? And I did go to test/ but there was no file named diagonal.jl.

[mcabbott]

No, you can literally paste my message in (the quoted part, for which github gives you a copy button). It is smart enough to remove the julia> prompts & ignore the answers.

This is the test file: https://github.com/JuliaLang/julia/blob/master/stdlib/LinearAlgebra/test/diagonal.jl#L777 . You should be able to copy & run bits of it, without running all of Julia's tests.