fix: Resolve worker teardown timeout by properly handling recursive scheduler promises

[phuongnd08]

Summary

Fixed a critical issue where Playwright E2E tests would complete successfully but the worker would timeout during teardown after 60 seconds. This was caused by floating promises in the scheduler's recursive chain.

Problem

The latest commit on main (7b1e53b6) broke the CI build with a worker teardown timeout:

• All 45 E2E tests pass successfully
• Last test completes at 09:00:35
• Worker teardown timeout occurs at 09:01:35 (exactly 60 seconds later)
• Error: Worker teardown timeout of 60000ms exceeded

Root Cause

In ClaudeTerminal.tsx, the scheduleNext() function had two issues:

1. Line 211: Called scheduleNext() recursively without returning the promise
2. Line 219: Called scheduleNext() without any promise handling

This created untracked promise chains that continued running even after tests completed, preventing the Playwright worker from cleaning up properly.

Solution

1. Added explicit Promise<void> return type to scheduleNext function for clarity
2. Changed line 213 to return scheduleNext() to properly chain the recursive promise
3. Used void operator on line 221 to explicitly mark the initial call as fire-and-forget while maintaining the promise chain internally

This ensures that when stopScheduler() is called (either explicitly or during test cleanup), the entire promise chain can resolve properly, allowing the Playwright worker to teardown without timing out.

Changes

• apps/desktop/src/renderer/components/ClaudeTerminal.tsx: Fixed promise handling in scheduler

Test Plan

• Local lint and type check passes
• CI E2E tests complete without worker teardown timeout
• Verify fix by running CI build 3 times consecutively

Evidence

CI logs from failed build (19359320911):

2025-11-14T09:00:35.9366332Z   ✓  45 [electron] › e2e/worktree-terminal-content-preservation.spec.ts:82:7
2025-11-14T09:01:35.9412859Z [31mWorker teardown timeout of 60000ms exceeded.[39m

🤖 Generated with Claude Code