Skip to content

cdshealpix: Add version 0.8.1 - #3224

Draft
luhenry wants to merge 1 commit into
mainfrom
cdshealpix
Draft

luhenry wants to merge 1 commit into
mainfrom
cdshealpix

Conversation

@luhenry

@luhenry luhenry commented Oct 14, 2026 •

Copy link
Copy Markdown
Member

Compiles the PyO3/maturin bindings to the pure-Rust cdshealpix HEALPix crate, one wheel per interpreter. Upstream publishes no riscv64 wheel.

Mirrors the Linux jobs of upstream's deploy_pypi.yml and the pytest run of test.yml.

Differs from upstream

  • cibuildwheel instead of a maturin publish loop - runs the test suite on each interpreter
  • No musllinux - numpy, a hard runtime dependency, has no musllinux riscv64 wheel

Matrix: cp312/cp313/cp314; no cp314t, upstream ships no free-threaded wheel.

Testing

  • same as upstream (--doctest-modules, benchmark module ignored), against the installed wheel
  • MPLBACKEND=Agg replaces upstream's root conftest.py, which only selects that backend

License: OK

Built on cp312 (riscv64 under QEMU); 136 passed.

Build the PyO3/maturin HEALPix extension per interpreter (cp312-cp314)
with cibuildwheel from the v0.8.1 tag, mirroring upstream's
deploy_pypi.yml Linux jobs and the pytest run of test.yml. No cp314t
(upstream ships none) and no musllinux (numpy has no musllinux riscv64
wheel).

luhenry commented Oct 14, 2026

Copy link
Copy Markdown
Member Author

check_commit_messages (job 115859831944) failed on check_no_ai_attribution, flagging a "Generated by Claude Code" footer in the PR description. Not a real issue with the current PR: create_pull_request auto-injects that footer, and it was stripped via update_pull_request right after creation — the live PR body has never had it since. The failed run just evaluated the pull_request "opened" webhook's frozen body snapshot (captured before the strip), which a re-run can't refresh since re-runs reuse the original event payload.

The edit fired a separate pull_request.edited run (workflow run 38566358252) carrying the corrected body; its check_commit_messages job is queued behind the backlog and should pass once it runs. No fix to push here — watching for that run to clear (same situation as #3221).


Generated by Claude Code

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant