mrhhsg opened a new pull request, #63578:
URL: https://github.com/apache/doris/pull/63578

   ### What changed
   
   Remove the extra `_alreay_eos` scanner state from the scan projection path.
   
   The final two-block eos handoff is now represented with the existing 
`_origin_block`:
   - if get_block() reports eos while both `_origin_block` and `_padding_block` 
contain data, return the current origin block with `eos=false`;
   - after projection succeeds, move the pending padding block into 
`_origin_block`;
   - on the next call, `_origin_block` is non-empty at entry, so return it with 
`eos=true`.
   
   ### Why
   
   `_alreay_eos` only duplicated state that can be derived from the block 
lifecycle, and it also carried a typo in the member name. Keeping the pending 
final block in `_origin_block` removes the separate boolean without changing 
the scanner output behavior.
   
   ### Validation
   
   - Formatted modified C++ files with `build-support/run_clang_format.py` 
using clang-format 16.
   - `ninja -C be/ut_build_ASAN -j 1 
src/exec/CMakeFiles/Exec.dir/scan/scanner.cpp.o`
   - `ninja -C be/ut_build_ASAN -j 1 
test/CMakeFiles/doris_be_test.dir/exec/scan/scanner_late_arrival_rf_test.cpp.o`
   - Attempted `./run-be-ut.sh --run 
--filter=ScannerProjectionTest.returns_padding_block_after_limit_eos_without_extra_flag:ScannerLateArrivalRfTest.applied_rf_num_advances_after_late_arrival`;
 the full BE UT build did not complete because the shared build machine hit 
process exhaustion: `ccache: error: Failed to fork: Resource temporarily 
unavailable`.
   
   ### Release note
   
   None
   


-- 
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]

Reply via email to