Gracefully error when using multidimensional indices

The following error is really hard to debug, particularly given the repeats 2 times which hides what the type of the input is.

Example)

julia> using AxisArrays

julia> aa = AxisArray(rand(10, 5), Axis{:time}(1:10), Axis{:name}([:a, :b, :c, :d, :e]))
10×5 AxisArray{Float64,2,Array{Float64,2},Tuple{Axis{:time,UnitRange{Int64}},Axis{:name,Array{Symbol,1}}}}:
 0.204148    0.792777   0.780203  0.222078   0.530742
 0.790793    0.684368   0.328124  0.475048   0.407242
 0.0292672   0.67966    0.540458  0.429185   0.694913
 0.932434    0.748893   0.831468  0.423212   0.496423
 0.127433    0.191755   0.404633  0.0733558  0.689208
 0.00480001  0.0853037  0.520147  0.29591    0.44223
 0.0359702   0.990168   0.636954  0.717789   0.832071
 0.0701029   0.94967    0.950602  0.2058     0.56855
 0.682146    0.285136   0.287865  0.97511    0.89489
 0.370766    0.569141   0.42673   0.917962   0.91706

julia> mask = iseven.(1:10)
10-element BitArray{1}:
 false
  true
 false
  true
 false
  true
 false
  true
 false
  true

julia> other = Array{Bool}(trues(5))
5-element Array{Bool,1}:
 true
 true
 true
 true
 true

julia> aa[mask, other']
ERROR: ArgumentError: index CartesianIndex{2}[CartesianIndex(1, 1), CartesianIndex(1, 2), CartesianIndex(1, 3), CartesianIndex(1, 4), CartesianIndex(1, 5)] not found
Stacktrace:
 [1] axisindexes at /Users/rory/.playground/share/simulation/depot/packages/AxisArrays/G6pZY/src/indexing.jl:302 [inlined]
 [2] axisindexes at /Users/rory/.playground/share/simulation/depot/packages/AxisArrays/G6pZY/src/indexing.jl:174 [inlined]
 [3] macro expansion at /Users/rory/.playground/share/simulation/depot/packages/AxisArrays/G6pZY/src/indexing.jl:354 [inlined]
 [4] to_index at /Users/rory/.playground/share/simulation/depot/packages/AxisArrays/G6pZY/src/indexing.jl:311 [inlined]
 [5] getindex at /Users/rory/.playground/share/simulation/depot/packages/AxisArrays/G6pZY/src/indexing.jl:117 [inlined] (repeats 2 times)
 [6] top-level scope at none:0

Reproduce the example in the issue, then inspect src/indexing.jl around the reported axisindexes, to_index, and getindex locations. Trace how the multidimensional indices become the reported CartesianIndex values; done means the same input produces a clear error that identifies the problematic input and indexing situation.