xianwenchen

My bad. I did not notice the difference. Shall I copy the content over to a new ticket for BinaryBuilder.jl and close this one?

Thanks, Yuyi. I'm just following the leads and trying to talk to people that may help resolve the issue. Right now it seems that the issue is in the setting...

Thank you. Yeah. It seems that it was that line that was causing the issue: https://github.com/JuliaBinaryWrappers/GDAL_jll.jl/blob/a242496e63a7656ee175d31d20bf025165567e07/src/wrappers/aarch64-linux-gnu-cxx11.jl#L461 .26 was chosen by the developers of the gdal library version 3.0.x. They switched...

Out of curiosity, can I use `ccall` to call my system's `libgdal.so.26` directly?

A ticket is opened at https://github.com/JuliaPackaging/BinaryProvider.jl/issues/198

Just to refine the answer from Freed-Wu, one needs to install the sci-visualization/gr package on Gentoo. As of today, the package is masked for amd64. Thus one has to unmarsk...

I forgot to say that I am using Julia 1.5.3 on Linux and 1.5.4 on Windows. Both produced same results.

@nalimilan, could you have a quick look at the issue and confirm or reject that it is a problem of GLM?

I also started a discussion on Discourse: https://discourse.julialang.org/t/can-someone-replicate-this-glm-problem-with-linear-regression-on-your-computer/60098 rafael.guerra found out that by subtracting 2000 from Year, the regression was estimated correctly. He wonders if the problem was due to...

Thank you. This indeed solved the issue. I will prepare a note for students so that they are aware. As I have commented on Discourse, I’m not sure if this...