Devdha commented on PR #3880: URL: https://github.com/apache/parquet-java/pull/3880#issuecomment-6434390621
Second Opus 5.5 review (via Claude Code) of the fix: **no blockers**, mergeable after normal review. Addressed the remaining should-fix items in `35e2079`: - flush test now asserts the target was flushed and contains bytes *before* `finish()` - added tests for invalid quality (decompression still works, compression rejected), zero/negative size bounds, and close-after-finish-failure idempotency - bomb tests assert the `exceeds maximum size` reason - documented that the generic `createInputStream(InputStream)` path buffers the decoded stream; Parquet uses the bounded overload - invalid quality now throws `BadConfigurationException` for consistency with `CodecFactory` - write/flush failures are recorded and reported on later `finish()`/`close()` Verification: `TestBrotliCodec` 11/11, `TestDirectCodecFactory#compressionCodecs` 1/1, `spotless:check` and `package` pass. Remaining: CI coverage on linux-aarch64, and a full `ParquetWriter`/`ParquetReader` round-trip with >4 MiB pages (manual pyarrow interop was checked). -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
