Repository navigation
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (1)
💤 Files with no reviewable changes (1)
📝 WalkthroughWalkthroughThe patcher now treats ChangesOpenClaw chat-send queue patch
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRsSuggested labels
Suggested reviewers
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
There was a problem hiding this comment.
🧹 Nitpick comments (1)
scripts/patch-openclaw-chat-send.js (1)
134-177: 🩺 Stability & Availability | 🔵 TrivialRun the targeted TUI correlation E2E before merge.
These unit tests cover the text rewrite, but this helper also affects live turn ordering and duplicate-send behavior in the sandbox image. Please run
openclaw-tui-chat-correlation-e2eon this branch before shipping. As per path instructions, this helper can affect “TUI message ordering, duplicate sends, and user/assistant turn correlation” and the recommended selective check isopenclaw-tui-chat-correlation-e2e.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@scripts/patch-openclaw-chat-send.js` around lines 134 - 177, Run the targeted openclaw-tui-chat-correlation-e2e before merging this patch, since the change in the chat send helper can affect live turn ordering, duplicate sends, and user/assistant correlation beyond the text rewrite covered by unit tests. Verify the updated logic in the patch-openclaw-chat-send.js helper, especially the resolvedQueue rewrite and the webchat/msteams condition, with the selective TUI correlation E2E check.Source: Path instructions
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@scripts/patch-openclaw-chat-send.js`:
- Around line 134-177: Run the targeted openclaw-tui-chat-correlation-e2e before
merging this patch, since the change in the chat send helper can affect live
turn ordering, duplicate sends, and user/assistant correlation beyond the text
rewrite covered by unit tests. Verify the updated logic in the
patch-openclaw-chat-send.js helper, especially the resolvedQueue rewrite and the
webchat/msteams condition, with the selective TUI correlation E2E check.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 86c65a9d-139c-4b0c-b707-3d30e9259cf5
📒 Files selected for processing (2)
scripts/patch-openclaw-chat-send.jstest/openclaw-chat-send-patch.test.ts
|
✨ Thanks for extending the OpenClaw chat.send patch to queue Teams turns with the same per-message followup behavior as webchat. This proposes a way to upgrade older webchat-only shims and adds regression coverage for fresh patching, reruns, and old-shim upgrade behavior. Related open issues: |
|
Thanks for putting this together. I’d like to hold this PR for now rather than merge it on the current OpenClaw baseline. After looking at #5851 again, the duplicate/stale Teams reply behavior looks likely to be coming from the OpenClaw runtime path. #5595 is already bumping NemoClaw to OpenClaw 2026.6.9 and updates the same If the issue still reproduces after #5595 is checked in, we can rebase this PR on top of the new OpenClaw patch shape and decide whether extending the queue-mode behavior to Thanks again for the quick patch and regression coverage. |
299b60a to
ed006f4
Compare
|
Rebased on current main, including the merged #5995 baseline. The patch test, build, typecheck, and lint/checks pass; leaving this parked for QA confirmation. |
Refs NVIDIA#5851 Signed-off-by: Deepak Jain <deepujain@gmail.com>
ed006f4 to
0f912e9
Compare
|
Rebased on current main and resolved the OpenClaw 2026.6.10 test-shape conflict while keeping the Teams chat.send regression. Patch test, build, typecheck, lint/checks, and the conditional scan pass; still parked for QA confirmation. |
|
Thanks @deepujain for supporting on this duplicate changes. After #5995 is merged, QA confirmed it's fixed and no longer repro. I'm closing this PR. |
Summary
Changes
sessionCtx.Provider === "webchat"to webchat ormsteams.Testing
npx vitest run test/openclaw-chat-send-patch.test.tsnpm run build:clinpm run typecheck:clinpm run source-shape:checkgit diff --checkEvidence it works
The regression test rewrites a freshly patched OpenClaw fixture back to the old webchat-only shim, reruns the patch script, and verifies it upgrades to the webchat/Teams guard without leaving the old provider check behind.
Refs #5851
Signed-off-by: Deepak Jain deepujain@gmail.com
Summary by CodeRabbit