Tamas Nagy

I suspect this is an issue with `Blink.jl`. Not sure TBH. ```julia using Compose #master using Gadfly #master c = compose(context(), rectangle()) ``` renders just fine in my hands.

Sorry I wasn't specific. In a browser window. That approach uses Gadfly's rendering to display the Compose context.

I just ran into this problem: ```julia julia> using AxisArrays julia> using Unitful: μm julia> img = AxisArray(ones(10, 10), Axis{:x}(1μm:1μm:10μm), Axis{:y}(1μm:1μm:10μm)); julia> img[Axis{:x}(atindex(-2μm .. 2μm, 5)), Axis{:y}(atindex(-2μm .. 2μm, 5))]...

@timholy any thoughts on the performance implications of returning an array with an eltype of `Union{T, Missing}`?

bump

This has always seemed super clunky to me ```julia for i in AxisArrays.axes(img, Axis{:x}()) ... end ``` > I collected images with non-uniform sampling time We do this a lot...

I think your last idea would work if the user has autoformatting enabled, but what if they don't? I'm not sure how pandoc itself actually handles it.

Haha though that's mostly Markdown's fault. Actually, I don't think so. Here's what I'm thinking. This is mostly a problem because the lines following the number are indented, which is...

Where are ink's settings? I can't seem to find them under the package screen in Atom.

It would be a great to have this feature. I had this problem with displaying pvalues where I would like to show 4 significant figures over several orders of magnitude.