Pluto.jl: GeoArrays working in REPL, not working in Pluto.jl

Hi,

I run into an issue using GeoArrays, (and ArchGDAL). When using these in the REPL all is working correctly. In Pluto.jl I get the following error on cell 2d7cfe40-eee1-11ea-32f4-8fff47db1e13 (see notebook below):

InitError: could not load library “[userdir].julia\artifacts\93a125a1f028ecc90665616e44072e4da65aea23\bin\libcurl-4.dll”

If I run the cell again the error becomes:

InitError: could not load library “[userdir].julia\artifacts\8ff6feb9234a4584dbdb04c7df64425d67e7a3d9\bin\libproj_7_0.dll”

3th time:

InitError: could not load library “[userdir].julia\artifacts\1a3f285b91b22d9081d1ec6a0af713214f300fd6\bin\libgdal-26.dll”

4th and 5th time:

InitError: could not load library “libgdal-26.dll”

6th time: No error on this cell! But the last error moved to the last cell.

System: Windows 10 Julia 1.5.1

Notebook:

### A Pluto.jl notebook ###
# v0.11.12

using Markdown
using InteractiveUtils

# ╔═╡ 206fc6b0-eee1-11ea-1e61-dd3fe339bdf8
begin
	import Pkg
	Pkg.activate()
	Pkg.Registry.update()
end

# ╔═╡ 2d7cfe40-eee1-11ea-32f4-8fff47db1e13
begin
	Pkg.add("GeoArrays")
	using GeoArrays
end

# ╔═╡ 7da3dfde-eee3-11ea-06ab-6bfb1ccbdcb9
begin
	fn = download("https://github.com/yeesian/ArchGDALDatasets/blob/master/data/utmsmall.tif?raw=true")
	geoarray = GeoArrays.read(fn)
end

# ╔═╡ Cell order:
# ╠═206fc6b0-eee1-11ea-1e61-dd3fe339bdf8
# ╠═2d7cfe40-eee1-11ea-32f4-8fff47db1e13
# ╠═7da3dfde-eee3-11ea-06ab-6bfb1ccbdcb9

I’m new to Julia/Pluto.jl so unsure if this is a bug or an error on my part.

About this issue

  • Original URL
  • State: closed
  • Created 4 years ago
  • Comments: 15 (6 by maintainers)

Most upvoted comments

I am still looking for someone to replicate this on windows

I just tried, and I could not reproduce on latest Pluto with the latest GDAL build, GDAL_jll v3.0.4+7. If you still have the problem, it’s maybe best to open a new issue on https://github.com/JuliaGeo/GDAL.jl, since Pluto is an unlikely culprit here.

There have been moments in the past that the GDAL build broke, in fact it was broken this week on some platforms until an hour ago. It should be getting more robust though, but it is a quite complex build. The good thing is that we have control over the environment, so if you get DLL errors from GDAL.jl, it’s best to submit an issue such that we can fix it.

I tried to reproduce this on linux and loading GeoArrays works in Pluto for me. You can use Pkg.status(mode=PKGMODE_MANIFEST) to get all dependencies. Have you tested, whether you can load the GDAL.jl package? I would suspect that it is a problem with the GDAL binary dependencies.