test(firestore): wait out the emulator's Listen-stream backoff

[tyler-reitz]

Why

test/firestore.test.tsx flakes at roughly 18% per run (11 of 60 iterations measured 2026-08-11). The Firestore emulator intermittently corrupts a Listen frame, grpc-js reads four body bytes as a length prefix, and the SDK answers the resulting RESOURCE_EXHAUSTED with a 60 second maximum backoff before self-healing.

This is an unresolved upstream emulator bug, firebase-tools#8654. Nothing to pin, so this survives it rather than preventing it.

What

Four waitFor calls in the two tests that read a nonexistent document get a 120s budget, and those two tests get a 150s vitest timeout. They are the only tests needing a live server round trip, so the only ones that can die; the rest are served from the local cache. waitFor polls at 50ms, so the larger ceiling costs nothing on a healthy run: 80-114ms per test with the change over five runs, 79-105ms without over three.

120s rather than 60s because a third party on the same upstream bug found 60s insufficient at least once and observed a self-heal at ~35s under 120s.

What is not verified

That this rescues a real occurrence. The desync has never reproduced on a dev machine (0 stalls in five fresh-emulator runs here). What is verified is the budget itself: a simulated 65s stall passes under it and fails under the default.

The evidence is CI after this lands: desyncs should keep appearing in logs while the flake stops. That is also the signal for closing #776.

The scoped cost: a genuine regression in these two tests now takes 120s to report instead of 1s. That is confined to the two tests this bug can kill.

Refs #776

[tyler-reitz]

Got a real occurrence in CI while this sat waiting, so the "not verified" caveat in the description is now partly answerable.

The test workflow has run nine times on this PR (18 job samples). Three attempts hit the desync; one of those landed on a vulnerable test and was rescued rather than failing:

GrpcConnection RPC 'Listen' stream error. Code: 8 RESOURCE_EXHAUSTED:
  Received message larger than max (1650553701 vs 4194304)
Using maximum backoff delay to prevent overloading the backend.
 ✓ test/firestore.test.tsx (15 tests) 12818ms
       ✓ returns undefined if document does not exist  9795ms

That test normally runs in ~90ms. Under the current ceilings it would have failed at 1000ms, or been killed by vitest at 5000ms. Recovery took about 10 seconds rather than the full 60, which matches the ~35s another project reported on the upstream issue and suggests the 120s budget has headroom.

18 of 18 jobs passed. On its own that is only p ≈ 0.03 against the 18.3% baseline flake rate, so the rescued stall is the persuasive part rather than the pass count.