CompatHelper: bump compat for "Showoff" to "1.0"

[github-actions]

This pull request changes the compat entry for the Showoff package from 0.2, 0.3 to 0.2, 0.3, 1.0.

This keeps the compat entries for earlier versions.

Note: I have not tested your package with this new compat entry. It is your responsibility to make sure that your package tests pass before you merge this pull request.

[rikhuijzer]

Could someone press the merge button on this one? Fixes #1535.

[rikhuijzer]

@Mattriks Thanks for merging. For me, this did solve the plotting issues. I checked that already before the reversal of Grisu (https://discourse.julialang.org/t/plotting-broken-gadfly-and-plots-backends-not-showing-axis-ticks-cant-plot-dataframes/60201/10). I'll see whether I can fix the tests.

[rikhuijzer]

The single failing test is still a Showoff issue. On Showoff v1.0.2:

julia> using Showoff

julia> showoff([100.0], :auto)
1-element Vector{String}:
 "100"

julia> showoff([0.0], :auto)
ERROR: InexactError: check_top_bit(UInt64, -9223372036854775481)
[...]

So, in summary. Showoff 0.3 was broken (but then not anymore because Kristoffer reversed a change) and Showoff 1.0 is also broken. I'll take it upstream.

[rikhuijzer]

It's a quite minor issue from upstream. Only when users pass one or more [0.0] elements, Gadfly will fail (which is exactly what happens in that issue106 test set). If they pass [0.0, 1.0], it does work.

I filed an issue upstream: JuliaGraphics/Showoff.jl#40.

[rikhuijzer]

For your information. The regression tests when switching over to Showoff to 1.0 show only one change for colorful_density:

master

devel

I guess this isn't too bad. Furthermore, I've submitted a PR to Showoff, so I hope they will fix that soon. Then, I expect that Gadfly's tests will pass again.

[rikhuijzer]

Could someone press the merge button on this one? Fixes #1535.

@Mattriks, what I said here was stated a bit in frustration yesterday. At that moment, I spent hours debugging something on my side which was caused by a broken Gadfly. I then spent hours trying to fix the upstream issue, only to find out that Gadfly already had a PR open for almost two months already which would have avoided the breakage. Plots.jl switched on the 5th of March (JuliaPlots/Plots.jl@08e3224).

Anyway, thanks for the merge. It should be easier to merge PRs now that CompatHelper will also run CI jobs (#1536).