Matrix division for `SMatrix{m, n, BigFloat}`

[hyrodium]

MWE:

julia> using StaticArrays

julia> M = @SMatrix rand(3,3)
3×3 SMatrix{3, 3, Float64, 9} with indices SOneTo(3)×SOneTo(3):
 0.0208401  0.45567   0.110421
 0.918376   0.652122  0.399424
 0.821561   0.418753  0.730901

julia> M / M
3×3 SMatrix{3, 3, Float64, 9} with indices SOneTo(3)×SOneTo(3):
 1.0          -0.0          -0.0
 1.24963e-16   1.0          -0.0
 8.33088e-17  -3.33235e-16   1.0

julia> N = @SMatrix rand(BigFloat, 3,3)
3×3 SMatrix{3, 3, BigFloat, 9} with indices SOneTo(3)×SOneTo(3):
 0.925296  0.56646   0.496179
 0.483511  0.669458  0.31921
 0.759881  0.802041  0.957391

julia> N / N
ERROR: setindex!() with non-isbitstype eltype is not supported by StaticArrays. Consider using SizedArray.
Stacktrace:
  [1] error(s::String)
    @ Base ./error.jl:33
  [2] setindex!
    @ ~/.julia/dev/StaticArrays/src/MArray.jl:39 [inlined]
  [3] macro expansion
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:66 [inlined]
  [4] _setindex!_scalar
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:46 [inlined]
  [5] setindex!
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:42 [inlined]
  [6] macro expansion
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:357 [inlined]
  [7] _setindex!
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:320 [inlined]
  [8] setindex!
    @ ~/.julia/dev/StaticArrays/src/indexing.jl:274 [inlined]
  [9] _solve
    @ ~/.julia/dev/StaticArrays/src/solve.jl:36 [inlined]
 [10] \
    @ ~/.julia/dev/StaticArrays/src/solve.jl:1 [inlined]
 [11] /(A::SMatrix{3, 3, BigFloat, 9}, B::SMatrix{3, 3, BigFloat, 9})
    @ LinearAlgebra ~/julia/julia-1.7.3/share/julia/stdlib/v1.7/LinearAlgebra/src/generic.jl:1152
 [12] top-level scope
    @ REPL[5]:1

This is because A/B uses setindex internally, and the method doesn't support non-isbitstypes. (x-ref: #883)
But is this error as expected?

I thought we can add a matrix division method for non-isbitstypes that does not use setindex.

[hyrodium]

Note that the division works fine for larger matrices:

julia> N = @SMatrix rand(BigFloat, 5,5)
5×5 SMatrix{5, 5, BigFloat, 25} with indices SOneTo(5)×SOneTo(5):
 0.406201   0.903827   0.657419   0.662293   0.30493
 0.591375   0.428022   0.799033   0.981537   0.514157
 0.175083   0.158292   0.136774   0.0813557  0.418362
 0.0178148  0.626641   0.0613683  0.17103    0.256846
 0.0244566  0.0628076  0.537964   0.33129    0.115918

julia> N / N
5×5 SMatrix{5, 5, BigFloat, 25} with indices SOneTo(5)×SOneTo(5):
 1.0          0.0   0.0          -0.0          -0.0
 4.77755e-78  1.0  -8.34278e-78  -6.25746e-78  -2.67633e-78
 0.0          0.0   1.0          -0.0          -0.0
 0.0          0.0   0.0           1.0          -0.0
 0.0          0.0   0.0          -0.0           1.0