j1wonpark opened a new pull request, #4338: URL: https://github.com/apache/amoro/pull/4338
## Why are the changes needed? Close #4337. Amoro depends on Iceberg 1.7.2. Among the Iceberg lines released since, 1.8.x is the only one that removes no engine module: `iceberg-spark-3.3` is dropped in 1.9.0, `iceberg-flink-1.18` in 1.10.0, and `iceberg-flink-1.19` in 1.11.0. Upgrading to 1.8.1 therefore keeps the full mixed-format engine matrix (Spark 3.3/3.4/3.5, Flink 1.18/1.19/1.20) intact while closing part of the gap to current Iceberg. Doing this hop on its own also keeps the API migration separate from the engine-support decisions that later upgrades will require. ## Brief change log - Bump `iceberg.version` from 1.7.2 to 1.8.1. - Migrate `RowDataFileScanTaskReader` in `amoro-mixed-flink-common-iceberg-bridge` from `FlinkAvroReader`, removed in Iceberg 1.8.0, to `FlinkPlannedAvroReader`. The class is a copy of the Iceberg class of the same name, so this mirrors the exact change Iceberg 1.8 made to its own copy (including `path().toString()` → `location()`); the only remaining deviations are the pre-existing mixed-format ones (`AdaptHiveFlinkParquetReaders`, shaded guava). No other production or test source changes are required. ## How was this patch tested? - Full `clean test-compile` passes against Iceberg 1.8.1 across all modules, including the `spark-3.3` and `spark-3.4` profiles. - `amoro-mixed-flink-common` + `amoro-mixed-flink-common-iceberg-bridge`: 314 tests pass (Kafka testcontainers included). - `amoro-mixed-hive`: 188 tests pass. - `spotless:check` and `checkstyle:check` pass. - Verified the bridge `RowDataFileScanTaskReader` against the Iceberg 1.8.1 source: after this change, the only remaining differences from the upstream class are the pre-existing mixed-format adaptations (`AdaptHiveFlinkParquetReaders`, shaded guava imports). - Built the shaded `amoro-mixed-flink-runtime-1.20` jar and confirmed it bundles the Iceberg 1.8.1 classes. -- 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]
