dbtsai opened a new pull request, #293:
URL: https://github.com/apache/spark-connect-rust/pull/293

   ### What changes were proposed in this pull request?
   
   Strengthen the shared release wheel recipe by passing `--locked` to Maturin 
and failing artifact upload when no wheel was produced. Constrain Maturin 
installation to `>=1.0,<2.0` in both wheel and sdist jobs, matching the 
existing `pyproject.toml` build-backend requirement, and document these build 
contracts.
   
   Follow-up to #165 and SPARK-60917. This is inherited release hardening; the 
original refactor retained these behaviors.
   
   ### Why are the changes needed?
   
   A manually selected release branch with an inconsistent lockfile can 
otherwise produce a wheel after Cargo silently updates its dependency graph. 
Requiring the checked-in lockfile makes the shared recipe enforce that contract 
independently of prior CI. The Maturin bound keeps direct CLI builds within the 
declared backend range, and an absent wheel should fail the build check instead 
of issuing an upload warning.
   
   ### Does this PR introduce _any_ user-facing change?
   
   No API or runtime change. Release wheel builds now require an up-to-date 
checked-in `Cargo.lock` and a produced wheel.
   
   ### How was this patch tested?
   
   - `actionlint` 1.7.12 passed the shared helper, wheel-check and release 
workflows; shell syntax and whitespace checks passed.
   - `cargo fmt --all --check` passed.
   - Genuine Cargo and Maturin 1.16.0 regression fixtures passed: a valid lock 
builds successfully, a stale lock fails with `--locked` and remains unchanged, 
and the unlocked control builds while rewriting the lockfile. All Cargo/Maturin 
processes used the serialized offline campaign wrapper and private 
tool/cache/target directories.
   - Structural checks verified that both Maturin CLI install ranges match 
`pyproject.toml`, artifact consumers keep their names, and no native/Python API 
or dependency inputs changed. Maturin's supported minimum v1.0.0 forwards 
`--locked` to metadata and compilation.
   - Genuine local Claude Code reviewed the entire follow-up diff after the 
second post-merge review; all findings are assessed in the review evidence.
   
   **Draft — required validation remains pending.** Repository library tests, 
integration-test compilation with and without `wasm-udf`, and Clippy were 
attempted but stopped before compilation because the required offline private 
cache lacks `prost`. The current project wheel builds, full release dry run, 
API introspection and official parity checks have not run for this patch. 
Existing PR #165 CI results are historical evidence, not validation of this 
follow-up. No checks are claimed as passing when blocked or unrun.
   
   ### Was this patch authored or co-authored using generative AI tooling?
   
   Generated-by: OpenAI Codex
   Reviewed-by: Claude Code 2.1.314 (`system.ai.claude-fable-5-1`)
   
   This pull request and its description were written by Isaac.
   


-- 
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]

Reply via email to