Stabilize diff pipeline TUI test

[warp-dev-github-integration]

Description

Stabilizes diff_pipeline_computes_added_lines_and_ghost_blocks from the required pre-fix baseline (453f4d30ca962406d7ac6004fa50cd4c879c51e3).

Root cause

The test started two asynchronous diff generations back-to-back:

1. reset_content emits ContentChanged and computes an empty diff because the reset content matches the new base.
2. apply_diffs immediately computes the intended replacement diff.

The test's one-shot subscription accepted the first DiffUpdated without identifying which generation completed. Background completion order can vary, and diff results are written without a version guard, so the empty reset result can be observed first or overwrite the edited result. The test then called expand_diffs and allowed only 100 scheduler yields for the downstream temporary-block layout. That bounded polling is also not a completion barrier: under scheduling variance, the intended diff and its ghost-layout action may not settle within 100 yields. The baseline reproduced with ghosts.len() == 0 on run 59 of a tight loop.

Fix

• Wait for the reset diff to finish before applying the edit, preventing the two diff generations from racing.
• Wait for the edited diff before expanding it.
• Replace the fixed 100-yield loop with LayoutInvalidated notifications plus a check of the actual ghost state.

This continues to exercise the real asynchronous diff and char-cell layout pipeline while synchronizing on the relevant generation and observable render state.

Linked Issue

No linked issue; this addresses the reported flaky Rust test.

• The linked issue is labeled ready-to-spec or ready-to-implement.
• Where appropriate, screenshots or a short video of the implementation are included below (especially for user-visible or UI changes).

Testing

• cargo test -p warp_tui diff_pipeline_computes_added_lines_and_ghost_blocks -- --nocapture

• Compiled exact test binary in a 1,000-run loop: 1,000/1,000 passed (baseline failed on run 59)

• cargo test -p warp_tui --lib (102 passed)

• cargo fmt --all -- --check

• cargo clippy -p warp_tui --tests -- -D warnings

• I have manually tested my changes locally with ./script/run (not applicable: test-only change)

Agent Mode

• Warp Agent Mode - This PR was created via Warp's AI Agent Mode

CHANGELOG-NONE

Co-Authored-By: Oz oz-agent@warp.dev

Conversation: https://staging.warp.dev/conversation/7cd7c094-1988-4dfd-bc65-e2f07d3e9c02
Run: https://oz.staging.warp.dev/runs/019f5d5a-e94e-76c2-944b-629b37da970d
Plans:

• Stabilize diff pipeline TUI test from pre-fix baseline

This PR was generated with Oz.