morningman opened a new pull request, #67828: URL: https://github.com/apache/doris/pull/67828
Cherry-picked from #67480 ### Cherry-pick notes Upstream #67480 was rebased onto #67594 (safe read fence for time-based incremental reads), which `branch-incremental-computation` does not carry. Conflicts in `TSOTimestamp.java`, `BindRelation.java`, `ExplainTableStreamPlanTest.java` and `TSOTimestampTest.java` were resolved by taking the PR side; all 11 files touched by the upstream squash commit are byte-identical to master after #67480. Two adaptations because #67594 is absent on this branch: 1. `BindRelation.parseTimestampRange`: the `endTimestamp` line still called the removed `composeFullTimestamp()`; switched to `composePhysicalTimestamp()` so the `@incr` range is `[startOf(startMs), startOf(endMs))`, matching the new BE `GE/LT` predicates. (On master this line came from #67594.) 2. `TimeBasedChangeVisibleWaiter` (+ its test): this branch still has the TSO-comparing waiter that #67594 replaced on master. Restored the waiter change from the PR's pre-rebase revision (`c1d7629524a`): the end bound is the exclusive start of the millisecond (`composePhysicalTimestamp`, formerly `composeEmptyCounterTSO`), the match uses `commitTSO < endTSO`, and the "read up to now" default maps query-start millisecond `M` to `startOf(M + 1)` so every logical counter within `M` is still awaited. This is the "Visibility waiting" behavior described in the upstream PR body. ### Verification - `be/src/exec/scan/olap_scanner.cpp`: syntax-only compile with the Release flags, clean. - FE UT (`run-fe-ut.sh --run`, fe-core main + test compiled against this branch): `OlapScanNodeTest` 12, `PhysicalPlanTranslatorTest` 17, `ExplainTableStreamPlanTest` 23, `TSOTimestampTest` 14, `TimeBasedChangeVisibleWaiterTest` 5 — 71 tests, 0 failures, 0 errors. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_018m1ARNXtGSWucJTy34uwe1 -- 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]
