Repository navigation
[Core] Fix issue #64692 by honoring client_processed_up_to also in the unordered actor task queue. - #67242
[Core] Fix issue #64692 by honoring client_processed_up_to also in the unordered actor task queue.#67242Pengkun-ZHU wants to merge 2 commits into
Conversation
There was a problem hiding this comment.
Code Review
This pull request introduces a mechanism to reject stale tasks in the UnorderedActorTaskExecutionQueue based on client-processed sequence numbers. It wraps queued tasks in a new SequencedTask struct containing a sequence number, tracks the highest client watermark (client_processed_up_to_) per concurrency group, and cancels tasks that are identified as stale. A unit test has also been added to verify this behavior. I have no feedback to provide.
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 4c8bc71. Configure here.
| struct SequencedTask { | ||
| int64_t seq_no; | ||
| TaskToExecute task; | ||
| }; |
There was a problem hiding this comment.
Missing Doxygen on new APIs
Low Severity
/** ... */ block comments using @ tags (@brief, and @param/@return where applicable).
Additional Locations (1)
Triggered by project rule: Bugbot Rules
Reviewed by Cursor Bugbot for commit 4c8bc71. Configure here.
… also in the unordered actor task queue. Main ideas are: - Remember the highest watermark received per concurrency group. - Carry each request’s RPC sequence number through every async wait and retry queue. - Check that sequence number against the latest watermark at the final execution decision. Signed-off-by: pzhu <q1091803103@gmail.com>


Description
Main ideas are:
Related issues
Fixes #64692
Additional information