After pushing a small change and tagging a new version of PSSFSS, Documenter is failing to run properly on the deployed documentation, though it works flawlessly on my local machine. When I look at the log file at https://github.com/simonp0420/PSSFSS.jl/runs/3036836943?check_suite_focus=true, I see that Documenter fails on each @example block. Here is a portion of the log file that describes the failure for the first such block in the documentation:
2021-07-10T17:45:11.4411764Z e[36me[1m[ e[22me[39me[36me[1mInfo: e[22me[39mSetupBuildDirectory: setting up build directory.
2021-07-10T17:45:13.5695128Z e[36me[1m[ e[22me[39me[36me[1mInfo: e[22me[39mDoctest: running doctests.
2021-07-10T17:45:15.8192441Z e[36me[1m[ e[22me[39me[36me[1mInfo: e[22me[39mExpandTemplates: expanding markdown templates.
2021-07-10T17:45:45.0030037Z e[33me[1m┌ e[22me[39me[33me[1mWarning: e[22me[39mfailed to run `@example` block in src/examples.md:67-91
2021-07-10T17:45:45.0031260Z e[33me[1m│ e[22me[39m```@example symmetric_strip
2021-07-10T17:45:45.0033011Z e[33me[1m│ e[22me[39musing Plots, PSSFSS
2021-07-10T17:45:45.0033995Z e[33me[1m│ e[22me[39mc = 11.802852677165355 # light speed [inch*GHz]
2021-07-10T17:45:45.0034853Z e[33me[1m│ e[22me[39mperiod = c # so the period/wavelength = freq in GHz
2021-07-10T17:45:45.0035595Z e[33me[1m│ e[22me[39mPy = period
2021-07-10T17:45:45.0036220Z e[33me[1m│ e[22me[39mLy = period/2
2021-07-10T17:45:45.0037016Z e[33me[1m│ e[22me[39mPx = Lx = Ly/10 # Infinite in x direction so this can be anything
2021-07-10T17:45:45.0037730Z e[33me[1m│ e[22me[39mNy = 60
2021-07-10T17:45:45.0038356Z e[33me[1m│ e[22me[39mNx = round(Int, Ny*Lx/Ly)
2021-07-10T17:45:45.0039111Z e[33me[1m│ e[22me[39msheet = rectstrip(;Px, Py, Lx, Ly, Nx, Ny, units=inch)
2021-07-10T17:45:45.0039842Z e[33me[1m│ e[22me[39mflist = 0.02:0.02:0.98
2021-07-10T17:45:45.0040859Z e[33me[1m│ e[22me[39msteering = (θ=0, ϕ=0)
2021-07-10T17:45:45.0041504Z e[33me[1m│ e[22me[39mstrata = [Layer()
2021-07-10T17:45:45.0042136Z e[33me[1m│ e[22me[39m sheet
2021-07-10T17:45:45.0042759Z e[33me[1m│ e[22me[39m Layer()]
2021-07-10T17:45:45.0043571Z e[33me[1m│ e[22me[39mresults_j = analyze(strata, flist, steering, showprogress=false,
2021-07-10T17:45:45.0044822Z e[33me[1m│ e[22me[39m resultfile=devnull, logfile=devnull);
2021-07-10T17:45:45.0045569Z e[33me[1m│ e[22me[39mp1 = plot(sheet)
2021-07-10T17:45:45.0046266Z e[33me[1m│ e[22me[39mp2 = plot(sheet, unitcell=true)
2021-07-10T17:45:45.0047104Z e[33me[1m│ e[22me[39mptitle = plot(title = "Symmetric Strip Triangulation",
2021-07-10T17:45:45.0047975Z e[33me[1m│ e[22me[39m grid = false, showaxis = false, xtick=[], ytick=[],
2021-07-10T17:45:45.0048758Z e[33me[1m│ e[22me[39m bottom_margin = -50Plots.px)
2021-07-10T17:45:45.0049563Z e[33me[1m│ e[22me[39mplot(ptitle, p1, p2, layout = @layout([A{0.09h}; [B C]]))
2021-07-10T17:45:45.0050475Z e[33me[1m│ e[22me[39msavefig("symstrip1.png"); nothing # hide
2021-07-10T17:45:45.0051170Z e[33me[1m│ e[22me[39m```
2021-07-10T17:45:45.0051758Z e[33me[1m│ e[22me[39m c.value =
2021-07-10T17:45:45.0055453Z e[33me[1m│ e[22me[39m MethodError: no method matching iterate(::Nothing)
2021-07-10T17:45:45.0057264Z e[33me[1m│ e[22me[39m e[0mClosest candidates are:
2021-07-10T17:45:45.0059357Z e[33me[1m│ e[22me[39m e[0m iterate(e[91m::Union{LinRange, StepRangeLen}e[39m) at range.jl:664
2021-07-10T17:45:45.0061934Z e[33me[1m│ e[22me[39m e[0m iterate(e[91m::Union{LinRange, StepRangeLen}e[39m, e[91m::Int64e[39m) at range.jl:664
2021-07-10T17:45:45.0063424Z e[33me[1m│ e[22me[39m e[0m iterate(e[91m::Te[39m) where T<:Union{Base.KeySet{var"#s79", var"#s78"} where {var"#s79", var"#s78"<:Dict}, Base.ValueIterator{var"#s77"} where var"#s77"<:Dict} at dict.jl:693
2021-07-10T17:45:45.0064248Z e[33me[1m│ e[22me[39m e[0m ...
2021-07-10T17:45:45.0065045Z e[33me[1m└ e[22me[39me[90m@ Documenter.Expanders ~/.julia/packages/Documenter/4JDQo/src/Expanders.jl:563e[39m
This happens for all of the @example blocks, of which there are quite a few. As I stated previously, this problem does not occur on my local build of the documentation. I checked, and it appears that the same version of Documenter is being used locally and for the deployed docs. Any help sorting this out would be greatly appreciated.