aajisaka opened a new pull request, #55753: URL: https://github.com/apache/spark/pull/55753
### What changes were proposed in this pull request? Update Dockerfile to fix build error: - Fix get-pip.py URLs to use versioned paths for Python 3.9 and PyPy 3.8 - Replace the gpg keyserver commands for the CRAN GPG key with a single `curl` command. gpg keyserver command is now not recommended way and the command doesn't work with `http_proxy` env var under proxy environment. - Remove `scipy` and `pandas` from the PyPy 3.8 pip install: neither can be installed in this environment. scipy >=1.8.0 fails due to a pythran/beniget/gast incompatibility; scipy <1.8.0 fails because numpy.distutils cannot locate Python.h at PyPy's non-standard header path. pandas has C++ extensions that trigger a PyPy 3.8 linker detection bug in setuptools (TypeError: 'NoneType' object is not subscriptable). Both scipy and pandas remain on the python3.9 install. ### Why are the changes needed? The build failure described in SPARK-56763 started to occur after https://github.com/apache/spark/commit/2a56312aeb1665b72c608e14926f5d69fd3a17bc. Before this commit, the build cache https://github.com/apache/spark/pkgs/container/spark%2Fapache-spark-github-action-image-cache/193506527?tag=branch-3.5 was used and the build was successful. - Before: https://github.com/apache/spark/actions/runs/24451638562/job/71441880373 - After: https://github.com/apache/spark/actions/runs/24562361796/job/71814012404 Without cache, the docker build fails due to a different error related to Python 3.8/3.9 EoLs. This commit fix the errors. ### Does this PR introduce _any_ user-facing change? scipy and pandas will be removed from PyPy 3.8 environment in the branch-3.5 base image ### How was this patch tested? Verified by running `docker build .` in `dev/infra/`. ### Was this patch authored or co-authored using generative AI tooling? Generated-by: Claude Code (Sonnet 4.6) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
