I think that I've narrowed this issue down to a sane minimal example, but I'm sure that I could improve on it. It may well be that what I'm doing is not yet intended to work.
The simple example
library(shiny)
install.packages ("ggvis")
library (ggvis)
mtcars %>% ggvis(~mpg, ~disp, fill = ~vs) %>% layer_points()
Produces this plot:
I get the same effect whether I use the redhat packaged rstudio-server, or the Centos 8 install described on the RStudio website (RStudio Server - Posit), rstudio-server-rhel-1.3.1056-x86_64.rpm
My simplest build is a vagrant box and a chrome browser. I'm installing:
R-shiny-1.4.0.2-2.fc32.noarch, R-Rcpp-devel-1.0.4.6-1.fc32.x86_64
My guess is that the installation of ggvis is breaking somewhere, but the installation appears clean.
Any pointers as to where I could be going wrong and/or who I need to point out the failure to?
tc