More modular hidedecoration settings for faceting

[piever]

There seems to be an edge case we do not cover, in that with

df = (
    sepal_length = 1 .+ rand(100), 
    sepal_width = 2 .+ rand(100),
    petal_length = 3 .+ rand(100),
    petal_width = 4 .+ rand(100)
)
xvars = ["sepal_length", "sepal_width"]
yvars = ["petal_length" "petal_width"]
layers = linear() + visual(Scatter)
plt = data(df) * layers * mapping(xvars, yvars, col=dims(1), row=dims(2))
draw(plt, facet=(linkxaxes=:none,))

we may actually wish to hide the inner axis labels, but keep the ticks.

[greimel]

Though this case x-axes should link by column, shouldn't they?

But you are right that more customisability is better.