What's changed
- Open large hive datasets without waiting on a full row count — hive-partitioned datasets now open immediately instead of blocking on a full scan to count every row first. (#44)
- Fix the nightly RPM build by sanitizing the
-devversion string. (#43) - Fix the v0.2.53 release-pipeline failures and clear the Node 20 deprecation warnings in CI. (#42)
Full changelog: v0.2.53...v0.2.54