Bounds-checking in triangular indexing branches

[jishnub]

After this PR, calls like the following would throw:

julia> U = UpperTriangular([1 2; 3 4])
2×2 UpperTriangular{Int64, Matrix{Int64}}:
 1  2
 ⋅  4

julia> U[1,0]
0

Currently, there's no bounds-checking on the indices if they don't correspond to stored values.

Also, this PR checks for compatible types in assigning to the diagonal of unit triangular matrices. This fixes issues like, e.g.:

julia> A = @SMatrix [1 2; 3 4]
2×2 SMatrix{2, 2, Int64, 4} with indices SOneTo(2)×SOneTo(2):
 1  2
 3  4

julia> M = fill(A, 2, 2)
2×2 Matrix{SMatrix{2, 2, Int64, 4}}:
 [1 2; 3 4]  [1 2; 3 4]
 [1 2; 3 4]  [1 2; 3 4]

julia> U = UnitUpperTriangular(M)
2×2 UnitUpperTriangular{SMatrix{2, 2, Int64, 4}, Matrix{SMatrix{2, 2, Int64, 4}}}:
 [1 0; 0 1]  [1 2; 3 4]
     ⋅       [1 0; 0 1]

julia> U[1,1] = 1
1

After this,

julia> U[1,1] = 1
ERROR: MethodError: Cannot `convert` an object of type Int64 to an object of type SMatrix{2, 2, Int64, 4}

However, this depends on JuliaLang/julia#58209 for the SizedArrays tests to pass, as oneunit currently doesn't work with a SizedArray.