WangGuangxin opened a new pull request, #13468: URL: https://github.com/apache/gluten/pull/13468
## What changes are proposed in this pull request? Move `CompressedInputStream` decompression timing from each `Read()` call to `RefillDecompressed()`. Sort shuffle reads row sizes and payloads separately, so timing every read adds clock reads even when the data is already decompressed. Timing refills removes that overhead while continuing to subtract blocking raw-input time. Add five native regression tests covering cached reads and timing, reads across compressed/decompressed chunks, concatenated frames and repeated EOF, empty frames and zero-length reads, and truncated frames. Cherry-picked from 3673833c27e9a53b1205748a5734649b33c6f3ac, with conflicts resolved against upstream main. The adaptation excludes unrelated test registrations present in the source commit's parent. ## How was this patch tested? - Passed `./dev/format-cpp-code.sh` with clang-format 15.0.7. - Passed `./.github/workflows/util/check.py header main --fix`. - Passed `git diff --cached --check` before committing. - Native tests were not run: CMake configuration is blocked by the missing Velox checkout (`scripts/setup-helper-functions.sh`); the local Docker daemon is unavailable. - Attempted `./dev/format-scala-code.sh`; Maven 3.9.16 download failed with HTTP 429. This patch changes no Java or Scala files. ## Was this patch authored or co-authored using generative AI tooling? Generated-by: Codex (GPT-6) -- 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]
