I’m afraid I don’t know much about debugging python-julia issues on Windows, and I can’t test this myself. However, one thing you might try is forcing PyCall (and, thus, PyPlot) to use its own self-contained Python version instead of whatever you may have on your system. Can you try running:
using PyCall
PyCall.python
if that outputs a version inside .julia/v0.6/Conda/deps/usr/bin/python
then you’re already using a self-contained Python version and this won’t help. But if it shows some other path on your computer, then you can try:
ENV["PYTHON"] = ""
Pkg.build("PyCall")
Pkg.build("PyPlot")
Failing that, if all you want to do is plot things using Plots.jl, you can just switch to a different backend. I like GR.jl
:
Pkg.add("GR")
using Plots
gr()
plot([1,2,3])