Fix bug in converting Zeros matrix to Diagonal

[dlfivefifty]

Fixes this bug on master:

julia> convert(Diagonal{Int,Vector{Int}}, Zeros{Int}(5,5))
ERROR: MethodError: no method matching Vector{Int64}(::Zeros{Int64, 2, Tuple{Base.OneTo{Int64}, Base.OneTo{Int64}}})

Closest candidates are:
  Array{T, N}(::FillArrays.AbstractZeros{V, N}) where {T, V, N}
   @ FillArrays ~/.julia/packages/FillArrays/eOEVm/src/FillArrays.jl:532
  Array{T, N}(::FillArrays.AbstractFill{V, N}) where {T, V, N}
   @ FillArrays ~/.julia/packages/FillArrays/eOEVm/src/FillArrays.jl:526
  Array{T, N}(::AbstractArray{S, N}) where {T, N, S}
   @ Base array.jl:673
  ...

Stacktrace:
 [1] convert(::Type{Vector{Int64}}, a::Zeros{Int64, 2, Tuple{Base.OneTo{Int64}, Base.OneTo{Int64}}})
   @ Base ./array.jl:665
 [2] Diagonal{Int64, Vector{Int64}}(diag::Zeros{Int64, 2, Tuple{Base.OneTo{Int64}, Base.OneTo{Int64}}})
   @ LinearAlgebra ~/Projects/julia-1.10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/diagonal.jl:10
 [3] convert(::Type{Diagonal{Int64, Vector{Int64}}}, A::Zeros{Int64, 2, Tuple{Base.OneTo{Int64}, Base.OneTo{Int64}}})
   @ FillArrays ~/.julia/packages/FillArrays/eOEVm/src/FillArrays.jl:540
 [4] top-level scope
   @ REPL[3]:1

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.91%. Comparing base (90ef3ce) to head (80b50c4).
Report is 4 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #389   +/-   ##
=======================================
  Coverage   99.91%   99.91%           
=======================================
  Files           8        8           
  Lines        1141     1141           
=======================================
  Hits         1140     1140           
  Misses          1        1           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jishnub]

Just a side note: convert(Diagonal{Int,Vector{Int}}, Zeros{Int}(5,5)) already works on julia v1.11+

[dlfivefifty]

Does that mean we should only add the special case for VERSION < v"1.11-"?

[jishnub]

Yes, this convert method would only be necessary on versions < v"1.11-". This conversion has now been added to LinearAlgebra. We may limit the method to older versions.