Bin Allocation and Support Selection w/ Geom.histogram

[choct155]

First and foremost, thank you for all the work you have done. Second, my issue concerns Geom.histogram. As detailed in this SO post, trying to plot the histogram on proportional data has resulted in a support that is too long and a single bar capturing all data on the [0,1] interval (so all of the data for proportional values). I have been advised that this may be a bug warranting an open issue. Am I doing something wrong?

[getzdan]

As, noted in the SO post, tracing the code leads to the function:
choose_bin_count_1d_discrete.
In it:

# minimum distance between two values
    mingap = zero(eltype(xs))
    for (a, b) in zip(1:length(xs_set)-1, 2:length(xs_set))
    mingap = mingap == zero(eltype(xs)) ? b - a : min(b - a, mingap)
end

seems to use the indices instead of the values, leading to problems further along.

[choct155]

Much appreciated! Also, enjoyed the JuliaCon talk.