Closed mathbunnyru closed 1 year ago
Hi! This is the friendly automated conda-forge-webservice.
I just wanted to let you know that I rerendered the recipe in conda-forge/r-devtools-feedstock#31.
Could you be more specific? This is a noarch
build, such that there is no compilation nor direct dependency on pinned libraries. So, a rebuild is unlikely to be the solution. Perhaps the metadata needs updating?
I got the error here: https://github.com/jupyter/docker-stacks/actions/runs/4101767597/jobs/7089277765
This is the Dockerfile which was built: https://github.com/jupyter/docker-stacks/blob/main/r-notebook/Dockerfile
I tried to investigate which package gives an error, and it seems that everything works fine if I exclude r-devtools
from the build.
Moreover, x86_64 build works just fine.
Perhaps the metadata needs updating?
I don't think so - we build docker images and always call mamba clean --all -f -y
after calling mamba install
.
I will investigate the problem (but not today); the problem might be in some of the downstream dependencies. It would be nice if mamba gave more information about what's going wrong by default.
Immediately, I think the problem is r-gert
is not built for linux-aarch64
. The latest versions have the dependency chain:
r-devtools -> r-usethis -> r-gert
I'll look into migrating it.
Thank you! 👍
There's a PR to start the process...but r-gert
has had problems with Windows builds recently. So, if it's not super urgent, I'll try working that first. If I can't get that working soon, we'll just temporarily skip Windows.
The linux-aarch64 builds for r-gert
went through without a problem. Give it a couple hours to get on Anaconda Cloud. Not sure if that will be sufficient, or if other packages also need updating.
@mathbunnyru there were a couple other dependencies that needed linux-aarch64 builds, but they're building now. I just checked a solve and everything seems to pull the latest versions now:
I'm going to close this and the PR out, but feel free to share any more feedback.
Thank you @mfansler. I restarted our pipeline: https://github.com/jupyter/docker-stacks/actions/runs/4122939982
Also, I will try to eliminate the differences between x86_64 and aarch64 images. If you can, please, take a look here: https://github.com/conda-forge/conda-forge-pinning-feedstock/pull/4078
Comment:
I get some version dependency bug with this package under Linux aarch64. I hope this will help.