[getappz]

Summary

Item #281 reported two distinct problems in the Telegram bot's inbound handling:

1. Plan/PR-approval cards were getting interleaved into unrelated conversational replies, because notify_human_gate/request_channel_approval (in src/supervisor/notify.rs) sent their Telegram API calls with no ordering against chat_channel::run_chat_turn's reply to the same chat.
2. When a chat's resumed agent session turned out to be stale and fell back to a brand-new zero-history session, the reply carried no indication of that — so a user's "Go ahead" / "Approved" answered by a session with no memory of the pending request read as the bot ignoring it, not as having genuinely forgotten.

Changes

• src/chat_channel.rs: chat_turn_lock is now pub(crate) so notify.rs can take the same per-chat lock run_chat_turn holds for its whole turn. Also adds prefix_session_reset_note, which prepends [my memory of this chat was reset -- replying fresh, with no earlier context] to a reply whenever the turn detected a stale resumed session (item chore: release #159's failure mode) and retried fresh.
• src/supervisor/notify.rs: notify_human_gate and request_channel_approval (shared by notify_pr_approval_gate and notify_plan_approval_gate) now take chat_turn_lock before sending, so a side-channel card can no longer land interleaved with an in-flight chat reply to the same chat.

Test plan

• cargo build --bin agentflare
• cargo test --bin agentflare chat_channel:: (new test: prefix_session_reset_note_prepends_context_loss_note)
• cargo test --bin agentflare supervisor:: (94 tests, all passing)
• cargo clippy --bin agentflare clean on both changed files
• cargo fmt --check clean

Note: a full-repo cargo test --bin agentflare run in this environment hits one unrelated failure, mcp_prompts::tests::optimize_review_returns_full_skill_body, caused by the sandbox mounting ~/.local/state read-only (that test's set_active() call has no test-time state-dir override) -- confirmed unrelated to this change by reproducing the same read-only-fs failure with a direct echo > ~/.local/state/... write, independent of any code change.

Summary by CodeRabbit

• New Features

  • Added a clear notice when a stale chat session is discarded and restarted, informing users that previous context was lost.

• Bug Fixes

  • Improved Telegram message sequencing so notifications and chat replies are delivered in the correct order without interleaving.
  • Ensured notification delivery continues even when an internal synchronization lock encounters an error.

[coderabbitai]

Warning

Review limit reached

Next included review available in 51 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available. Your 60 included PR review attempts over the past 7 days set your current allowance at 1 review per hour.

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Essentials

Run ID: 6aa435c9-0141-4031-80be-05efba601d0d

📥 Commits

Reviewing files that changed from the base of the PR and between c31c464 and 1ec52b2.

📒 Files selected for processing (2)

• src/chat_channel.rs
• src/supervisor/notify.rs

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

---

_{Comment @coderabbitai help to get the list of available commands.}

[getappz]

Tracking this PR for automated review (flared:51bb8de6c33b).