jk-shah opened a new pull request, #41436: URL: https://github.com/apache/beam/pull/41436
Adds `beam sql` to the experimental `beam` command introduced in #40968 and #41040. `beam sql` launches the Apache Beam SQL shell (`BeamSqlLine`) with the same flags, shaded uber-JAR build (including `delta` and `iceberg` connector parity from #40958 / #41090), and `~/.beam/cache/` + `~/.beam/maven-wrapper/` layout as `scripts/beam-sql.sh`, so cached JARs are shared between both entry points without rebuilding: ```bash beam sql beam sql --io csv --io parquet beam sql -e "SELECT 1 + 1 AS two;" beam sql --runner dataflow --io google-cloud-platform beam sql --list-versions beam sql --list-ios beam sql --list-runners ``` ### Key details - **Cross-platform & prerequisite checks:** Verifies a working Java 11+ runtime (`JAVA_HOME` or `PATH`, including `java.exe` on Windows and catching macOS's `/usr/bin/java` stub and Java 8 with exit code `4`), invokes `MavenWrapperMain` directly via `java` rather than `sh`/`mvnw`, and queries Maven Central via Python `urllib` (no `curl`, `sed`, or GNU `sort -V` required). - **Cache & network safety:** Uses cross-process file locks (`.maven-wrapper.lock`, `<jar>.lock` with `O_NOFOLLOW`), `zipfile.is_zipfile` validation (rebuilding truncated/corrupt cached JARs automatically), atomic `tempfile.mkstemp` + `os.replace` writes, HTTPS-only redirect enforcement, and `SNAPSHOT`-only snapshot repository gating. -- 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]
