fix(mcp): catch RuntimeError when canceling tasks on closed event loop during /exit

[liuhao1024]

What does this PR do?

Catches RuntimeError when canceling MCP server tasks on a closed event loop during /exit. When calling /exit, the shutdown sequence closes the asyncio event loop, but MCP tasks still running in _wait_for_reconnect_or_shutdown call t.cancel() → call_soon(), causing RuntimeError: Event loop is closed exceptions that are printed as "Exception ignored" noise.

The fix moves t.cancel() inside the try block and catches RuntimeError alongside CancelledError. This silences the noise without changing functional behavior (the exceptions were already being caught and ignored).

Related Issue

Fixes #60197

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)

Changes Made

• tools/mcp_tool.py: Move t.cancel() inside the try block and catch RuntimeError alongside CancelledError in two locations (line 1837 and 1882)

How to Test

1. Run the MCP tests: python -m pytest tests/tools/test_mcp_tool.py -q
2. All 203 tests pass
3. Observed result: calling /exit on an MCP-enabled session no longer emits "Exception ignored" RuntimeError noise

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

For New Skills

• N/A

[alt-glitch]

This was generated by AI during triage.
Duplicate of #60032 (earliest open, identical fix at the same two t.cancel() sites: cancel inside try + catch RuntimeError). Same mechanism as #60213; #60104 is a different approach (drain parked tasks before loop.close()). All target #60197.

[liuhao1024]

Thanks @alt-glitch for the duplicate detection. After comparing the diffs:

• fix(mcp): catch RuntimeError when cancelling tasks on a closed event loop #60032 (by @B416-JAFLY) was created ~6 hours earlier (2026-07-07T06:18:03Z)
• Both PRs fix the same issue ([Bug] RuntimeError: Event loop is closed during /exit (MCPServerTask.shutdown) #60197) by catching RuntimeError when canceling MCP tasks on a closed event loop
• Both modify the same two locations in tools/mcp_tool.py (line 1837 and 1882)
• fix(mcp): catch RuntimeError when cancelling tasks on a closed event loop #60032's approach is more conservative: wraps t.cancel() in its own try/except, preserving the original await t exception handling structure
• This PR's approach is more aggressive: moves t.cancel() inside the existing try block and changes the exception tuple to include RuntimeError

Since #60032 is earlier and provides the same fix with minimal code changes, maintainers should merge #60032 instead. Closing this PR as a duplicate.