alitrack commented on PR #964:
URL: https://github.com/apache/hudi-rs/pull/964#issuecomment-6568536078

   Following up on the Windows CI failures — resolved, and I want to be clear 
about what each fix covers.
   
   **`rust-tests-windows` build failure (protoc):** the `--all-features` run 
pulls in lance (via hudi-core's lance feature), whose `prost-build` build.rs 
needs `protoc`, which the Windows runner doesn't preinstall. `7bf11aa` installs 
it via `choco install protoc` (same pattern as `release.yml:86`).
   
   **Two pre-existing Windows test incompatibilities (unrelated to the #445 
forward-slash fix):** re-enabling the Windows job surfaces these; both fail 
**identically on unmodified `origin/main`**, so they are not regressions from 
this PR:
   
   - `storage::util::parse_valid_uri_in_various_forms` — its bare POSIX 
root-relative input `/foo/` is not a valid directory path on Windows (no drive 
letter), so `Url::from_directory_path` rejects it. I split that single case 
into its own `#[cfg(not(windows))]` test; the scheme-bearing cases 
(`file:/foo/`, `file:///foo/`, `hdfs`, `s3`, `gs`, `wasb`) still run on Windows.
   - `timeline::instant::test_create_instant_using_local_timezone` — fixes the 
offset via the POSIX `TZ` env var, which Windows ignores. Gated 
`#[cfg(not(windows))]`.
   
   Linux/macOS CI is unaffected by either gate.
   
   Local verification of the #445 fix itself: WSL `cargo test -p hudi-core 
--lib` = 1703 passed / 0 failed; Windows `datafusion` subset = 61 passed, 2 
failed (the two pre-existing cases above). The full `--all-features 
--workspace` Windows run is what this re-enabled CI job will now exercise 
end-to-end.
   


-- 
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]

Reply via email to