chat-open bench: keep the click reply's clicked flag (follow-up of #28940)
Summary
This is the follow-up from the review of #28940 at b7eb1d41 (receipt: https://github.com/manaflow-ai/cmux/pull/28940#issuecomment-6542437815). That review was SHIP, with one P2 follow-up.
The chat-open bench's click_subagent copied the app's debug.agent_pane click reply, then overwrote its clicked boolean with the subagent's name. So a bench artifact could not show a press that didn't happen. The reply now keeps the app's clicked and error fields as they came back, and records the subagent it clicked under subagent. The test fake now answers with clicked: true, as the app does, and the test checks that the flag survives.
Testing
python3 scripts/cmux-next/tests/test_chat_open_bench.py: all 16 tests pass.
Observed / Not run
- Observed: the unit tests above, and the app's click reply shape in
DebugAgentPaneClick.swift("clicked": .bool(pressed)). - Not run: a fleet bench at this head.
Changelog
none
Agent session: agent session
š¤ Generated with Claude Code
![View with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v3-3x.png)
![Autofix with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark-3x.png)
Need help on this PR? Tag @codesmith-bot with what you need.
Summary by cubic
Fixes the chat-open bench's click_subagent so bench artifacts keep the app's real clicked flag.
Previously, the function overwrote the reply's clicked boolean with the subagent's name, so an artifact could show a press that didn't happen. It now records the subagent under subagent and leaves clicked and error untouched. The test fake now returns clicked: true like the app does, and the test verifies the flag survives.