Doctests failing due to printing changes in 1.7

[fchorney]

https://github.com/invenia/AxisSets.jl/runs/4876807034?check_suite_focus=true

[iamed2]

A cursory look shows some of these related to how a Matrix with one column is displayed.

In 1.6:

julia> print(reshape([7.0; 8.0; 9.0], 3, 1))
[7.0; 8.0; 9.0]
julia> print([7.0; 8.0; 9.0])
[7.0, 8.0, 9.0]

Note that this is not round-trippable; both evaluate to a Vector

In 1.7:

julia> print(reshape([7.0; 8.0; 9.0], 3, 1))
[7.0; 8.0; 9.0;;]
julia> print([7.0; 8.0; 9.0])
[7.0, 8.0, 9.0]

this is round-trippable.

Some relate to Set ordering when printing, which can change when the RNG changes.

I'm not sure, but I think the last one is related to the printing of errors, where now failures can print a diff. It's hard to tell though as that's kind of meta.

[arnaudh]

Fixed with #70