chenjian2664 opened a new pull request, #10815: URL: https://github.com/apache/paimon/pull/10815
### Purpose Remove classes that no code in the repository uses anymore. Each module's removal is a separate commit, so it can be reviewed or reverted on its own. Two kinds of classes are removed: - Classes that nothing references, not even tests, docs, Python code or SPI service files. - Classes that only their own tests used, because an earlier refactor removed their last production caller. Their tests are removed with them. Removed per module: - **common**: `CastedVector`, `SortContext`, `BinaryRowDataUtil`, `BooleanArrayList`, `ListDelimitedSerializer`, `MapBuilder`, `ReuseByteArrayOutputStream`, `SerializablePredicate`, plus `LocalFileUtils`, `LongTripleArrayList`, `ObjectUtils` and `ConstantGlobalIndexReader`, whose last callers were removed earlier. `GlobalIndexScanner` now merges readers instead of padding shorter indexes, so the constant padding reader is no longer needed. - **core**: `RowIdIndexFieldsExtractor`, `IntObjectSerializer`. - **format**: `AvroSchemaTypelessReader`, `JodaConverter` and the Parquet `position` package. `JodaConverter` was the only user of the optional `joda-time` dependency, so that dependency is removed too. - **cdc**: `CdcRecordPartitionKeyExtractor`, and `AliyunFieldParser`, which `AliyunRecordParser` stopped using once it extracted types itself. - **ci**: `ScalaSuffixChecker`, which no CI script or build configuration runs. The Maven `dependency:tree` parsing that only it used is removed too: `DependencyTree`, the tree parsing in `DependencyParser`, and the scope and optional fields of `Dependency`. - **flink**: `RowDataKeyAndBucketExtractor` (unused since the sink moved to `InternalRow`), `NumberSequenceRowSource` (no longer built by aggregate push down), and the legacy `SinkFunction` shim in `paimon-flink1-common` (unused since the log system code was removed). - **spark**: the `SQLHelper` trait. Some classes are unused here but intentionally kept: public or deprecated API, compatibility shims, the Arrow native read/write extension points, Vortex JNI bindings that mirror the upstream native library, Flink CDC pipeline utilities used by the Flink CDC project, and recently added building blocks that are not wired in yet. ### Tests - Tests that only covered a removed class are removed with it. - `GlobalIndexEvaluatorTest` no longer tests padding shorter indexes with a constant reader, which production code no longer does. It still covers how the evaluator ANDs readers that cover different row ranges. - `CdcRecordKeyAndBucketExtractorTest` now compares the CDC extractor against `FixedBucketRowKeyExtractor` directly. That is the class the removed `RowDataKeyAndBucketExtractor` wrapped. - `TestCdcEventParser` falls back to an empty record list inline instead of calling `ObjectUtils`. -- 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]
