Fix isapprox for TensorMaps with different spaces

[Yue-Zhengyuan]

This PR improves isapprox for TensorMaps: when two TensorMaps have different spaces, isapprox should directly return false. Previously such a check is missing and a SpaceMismatch error will be raised.

[Jutho]

I think the current behavior is in line with how isapprox works on matrices:

julia> isapprox(rand(3,4), rand(5,6))
ERROR: DimensionMismatch: a has size (3, 4), b has size (5, 6), mismatch at dim 1
Stacktrace:
 [1] throw_promote_shape_mismatch(a::Tuple{Base.OneTo{…}, Base.OneTo{…}}, b::Tuple{Base.OneTo{…}, Base.OneTo{…}}, i::Int64)
   @ Base ./indices.jl:135
 [2] promote_shape
   @ ./indices.jl:196 [inlined]
 [3] promote_shape
   @ ./indices.jl:188 [inlined]
 [4] -(A::Matrix{Float64}, B::Matrix{Float64})
   @ Base ./arraymath.jl:7
 [5] isapprox(x::Matrix{…}, y::Matrix{…}; atol::Int64, rtol::Float64, nans::Bool, norm::typeof(LinearAlgebra.norm))
   @ LinearAlgebra ~/.julia/juliaup/julia-1.12.5+0.aarch64.apple.darwin14/Julia-1.12.app/Contents/Resources/julia/share/julia/stdlib/v1.12/LinearAlgebra/src/generic.jl:1985
 [6] isapprox(x::Matrix{Float64}, y::Matrix{Float64})
   @ LinearAlgebra ~/.julia/juliaup/julia-1.12.5+0.aarch64.apple.darwin14/Julia-1.12.app/Contents/Resources/julia/share/julia/stdlib/v1.12/LinearAlgebra/src/generic.jl:1981
 [7] top-level scope
   @ REPL[1]:1
Some type information was truncated. Use `show(err)` to see complete types.

[Yue-Zhengyuan]

Let me ask if LinearAlgebra.jl does this intentionally: JuliaLang/LinearAlgebra.jl#1624

[lkdvos]

As I mentioned in that issue there, I am not a huge fan of this, and would prefer if this just throws an error. My main objection is that the definition of isapprox(x, y) implies norm(x, y) > tol if !isapprox is true, which would no longer be the case here.