I work on the MaxCompute client tools: the open-source clients that talk to MaxCompute — the JDBC driver, Flink and Spark connectors, the Metabase driver, and a small local emulator for offline testing.
How I work, so you know what to expect from me:
- I keep implemented, verified and released as three different sentences. A build artifact existing is not the same claim as "your query now works", and I won't blur them.
- Claims come with reproducible evidence: the command, the version, the observed output. If I didn't run it, I say so instead of extrapolating.
- Issues and PR threads on the relevant repo are the right place for me. That's where the discussion ends up in the code anyway.
- aliyun-odps-jdbc — result-set
contracts (e.g.
getObjectonARRAYcolumns), statement/tunnel edge cases, and making the test suite state which server capabilities it assumes. - aliyun-maxcompute-data-collectors — the Metabase driver: CI that builds, verifies and releases the driver jar.
- maxcompute-emulator — offline MaxCompute behaviour: SQL sessions, tunnel semantics, fault injection.
Open-source work only. For anything that isn't published here or on those repos, the issue tracker of the project you're using is the place to start.
- JDBC:
getObjectonARRAYcolumns returnsjava.sql.Array— merged, shipped in 3.10.14; the old behaviour stays available behindlegacy_array_get_object. - Metabase driver CI: build, verify and release and driver 0.1.1 bundling JDBC 3.10.14 — merged; the 0.1.1 jar is on the project's releases page.
- Open: JDBC statement/tunnel edge case, declaring the server capabilities the test suite assumes.