test: stabilize flaky widget-tag streaming assertion (#3314)

[bolichen97]

test: stabilize flaky widget-tag streaming assertion

Problem

MochiChatPanel.coverage.test.tsx > ChatPanel streaming footer > drops a half-arrived widget tag rather than showing its markup fails intermittently in CI while passing locally. Observed on PR #3302, run 31710742636, shard Frontend Tests (4) at head cba9e9778 — a PR whose diff cannot reach this test's code paths.

Why it matters

A shard that fails on a file the diff cannot reach costs every unrelated PR a re-run and trains reviewers to distrust red CI.

Fix (symptom → root cause → change)

• Symptom: the test fails only on loaded 2-core CI runners, never locally.
• Root cause: after the stream mock emits content containing a partial <mcwidget tag, the test awaits the positive assertion (findByText('Building it now'), which waits) but then immediately fires the negative assertion with a synchronous queryByText(/mcwidget/). The streamed-state commit that strips the partial tag is asynchronous; on a slow runner the assertion runs before React flushes, still sees the pre-strip markup, and fails.
• Change: wrap the negative assertion in waitFor — await waitFor(() => expect(screen.queryByText(/mcwidget/)).not.toBeInTheDocument()) — so the assertion polls until React commits. This is the determinism-conventions remedy for this flake class (poll, never sleep/rerun/relax) and the same pattern this file already uses for its other post-async negative assertion ("replaces the streamed text with the committed message").

The assertion semantics are unchanged: if the panel ever rendered the partial widget markup persistently, waitFor times out and the test still fails.

Tests

This PR is itself a test change. MochiChatPanel.coverage.test.tsx passes 82/82; full frontend suite 17990 passed / 0 failed on the branch; npx tsc -b clean.

Manual verification

N/A — unit coverage sufficient; the change is confined to one test assertion.

Screenshots

N/A — no user-visible UI change.

Closes #3314

[github-actions]

UX Review (Fable 5) — ✅ PASS

Advisory UX-level review of 2c405b23e7b97c41c42cde6e743d5ff14c345215 — updated in place on each push; does not block merge.

UX-Verdict: PASS

Test-only change — one assertion wrapped in waitFor; no user-facing strings, components, or rendered pixels are touched.

[UX-REVIEWED] 2c405b2

[github-actions]

GPT 5.6 Review — ✅ no blocking findings

GPT 5.6 completed its review of 2c405b23e7b97c41c42cde6e743d5ff14c345215 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 2c405b2

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 2c405b23e7b97c41c42cde6e743d5ff14c345215: <one-sentence reason>

[github-actions]

Opus 4.8 Review — ✅ no blocking findings

Reviewed 2c405b23e7b97c41c42cde6e743d5ff14c345215 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 2c405b2

Verdict parsed from the review's SHA-scoped output markers for commit 2c405b23e7b97c41c42cde6e743d5ff14c345215.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 2c405b23e7b97c41c42cde6e743d5ff14c345215: <one-sentence reason>

[github-actions]

Design Review (Fable 5) — ✅ PASS

Advisory design-level review of 2c405b23e7b97c41c42cde6e743d5ff14c345215 — updated in place on each push; does not block merge.

Design-Verdict: PASS

Root-cause fix for a diagnosed CI race, using the repo's documented flake remedy and the file's own existing pattern; semantics preserved.

[DESIGN-REVIEWED] 2c405b2

[iamwhatever]

Tier 1 auto-approve: test stabilization (1 file). Criteria: no conflict, no requested changes, no security surface, AI reviewers green. Category: flaky test fix — stabilizes widget-tag streaming assertion timing (#3314 follow-up).