Repository navigation
knowledge(vllm-omni): 1 change(s) from refine - #354
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: aabc20bcce
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
|
||
| - [native sender-only admission 与 opt-in first-chunk express](rules-first-chunk-express.md)。 | ||
| - [Async-chunk prewarm 规则](rules-async-chunk-prewarm.md) | ||
| - [Scheduler：duplex 流式 context overflow 规则](rules-duplex-overflow.md) |
There was a problem hiding this comment.
Linking this page only from _index.md makes the new rule unreachable in the required review flow: direct_review_plan() for the referenced PersonaPlex overflow change returns the scheduler rules.md quick map, which contains no overflow entry, while KnowledgeDocs.related() deliberately excludes type: rule pages and returns no related document. Because Direct then stops index navigation, reviewers never receive this rule; add it to the scheduler rules.md Direct map rather than only the index.
AGENTS.md reference: knowledge/AGENTS.md:L16-L16
Useful? React with 👍 / 👎.
|
|
||
| ## VLLM-OMNI-PR8683-DUPLEX-OVERFLOW-FENCE — 流式 context overflow 必须走 duplex 失败路径，overflow fence 只能由 FIFO abort 释放 | ||
|
|
||
| - 触发：duplex/PersonaPlex streaming 会话的 prompt 达到或超过 `max_model_len`；或修改 `vllm_omni/core/sched/omni_ar_scheduler.py` 的 streaming context overflow 路径（`_overflowed_streaming_requests` fence、`release_overflowed_requests`）、`vllm_omni/engine/stage_engine_core_proc.py` 的 eager/FIFO abort 双队列处理，或 `vllm_omni/engine/duplex_orchestrator.py` 的终端错误路由。 |
There was a problem hiding this comment.
For changes limited to stage_engine_core_proc.py or duplex_orchestrator.py, these FIFO-abort, terminal-routing, and session-policy requirements are stored under Scheduler even though this directory's own _index.md assigns engine orchestration to Serving and the Serving index owns vllm_omni/engine/ and duplex orchestration. Owner-scoped Direct routing can therefore finish at Serving without loading this page; split the engine requirements into the existing Serving duplex/abort rules and keep only the scheduler fence contract here.
AGENTS.md reference: AGENTS.md:L12-L17
Useful? React with 👍 / 👎.
Generated automatically by the Copilot knowledge service (changeset
vllm-omni-refine-b7f5936dd332).repos/vllm-omni/components/scheduler/rules-duplex-overflow.mdGenerator:
zcode:GLM-5.3· judge:codex:gpt-6-sol:medium.The publisher merges it only if its local gate passes on the exact merge result.