Support Symmetric Matrices

The problem is not resolved by defining Base.convert(::Type{Symmetric}, x::Array{Float64, 2}).

I believe the problem is the non-commutativity of Symmetric and BlockArray methods.
Specifically, I think the sub-block types are being defined as Symmetric in this line and then setindex! complains when it tries to assign an Array{T, 2} to the blocks.

At least that's my educated guess. Folks more familiar with what those functions do will know more.

julia> using BlockArrays, LinearAlgebra

julia> Symmetric(BlockArray(rand(5, 5), [3, 2], [2, 3]))
5×5 Symmetric{Float64,BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}} with indices 1:1:5×1:1:5:
 0.80611   0.980725  │  0.396233  0.218845  0.126326
 0.980725  0.519793  │  0.370195  0.88799   0.996251
 ────────────────────┼──────────────────────────────
 0.396233  0.370195  │  0.344366  0.171352  0.478706
 0.218845  0.88799   │  0.171352  0.789426  0.58824 
 0.126326  0.996251  │  0.478706  0.58824   0.565382

julia> BlockArray(Symmetric(rand(5, 5)), [3, 2], [2, 3])
ERROR: MethodError: Cannot `convert` an object of type Array{Float64,2} to an object of type Symmetric{Float64,Array{Float64,2}}
Closest candidates are:
  convert(::Type{#s664} where #s664<:Symmetric, ::Union{Hermitian, Symmetric}) at /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.4/LinearAlgebra/src/symmetric.jl:193
  convert(::Type{T}, ::T) where T<:AbstractArray at abstractarray.jl:14
  convert(::Type{T}, ::Factorization) where T<:AbstractArray at /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.4/LinearAlgebra/src/factorization.jl:55
  ...
Stacktrace:
 [1] setindex! at ./array.jl:828 [inlined]
 [2] setblock! at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:400 [inlined]
 [3] setindex! at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/abstractblockarray.jl:203 [inlined]
 [4] macro expansion at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:184 [inlined]
 [5] macro expansion at ./cartesian.jl:64 [inlined]
 [6] macro expansion at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:181 [inlined]
 [7] BlockArray{Float64,N,R,BS} where BS<:Tuple{Vararg{AbstractUnitRange{Int64},N}} where R<:(AbstractArray{#s13,N} where #s13<:AbstractArray{Float64,N}) where N(::Symmetric{Float64,Array{Float64,2}}, ::Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}) at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:179
 [8] BlockArray{Float64,N,R,BS} where BS<:Tuple{Vararg{AbstractUnitRange{Int64},N}} where R<:(AbstractArray{#s13,N} where #s13<:AbstractArray{Float64,N}) where N(::Symmetric{Float64,Array{Float64,2}}, ::Array{Int64,1}, ::Array{Int64,1}) at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:172
 [9] BlockArray(::Symmetric{Float64,Array{Float64,2}}, ::Array{Int64,1}, ::Array{Int64,1}) at /Users/glenn/.julia/packages/BlockArrays/Iqf2f/src/blockarray.jl:175
 [10] top-level scope at REPL[199]:1
 [11] eval(::Module, ::Any) at ./boot.jl:331
 [12] eval_user_input(::Any, ::REPL.REPLBackend) at /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.4/REPL/src/REPL.jl:86
 [13] run_backend(::REPL.REPLBackend) at /Users/glenn/.julia/packages/Revise/tV8FE/src/Revise.jl:1165
 [14] top-level scope at none:0

Reproduce the failure with Symmetric and BlockArray using the examples in the issue. Read src/blockarray.jl around lines 172-184 and 400, then trace how blocks are created and assigned; done means both shown constructions work without the conversion error and the relevant behavior is covered by tests.