[andrewkrippner]

Summary

• After Auto Takeoff, onSpaceWrite writes to 100+ spaces one at a time. Firestore then delivers that many snapshots in one task, and each one called notify(). That gave one synchronous useSyncExternalStore commit per snapshot, and React 19 threw #185 (Maximum update depth) once 50+ commits ran back to back while an effect kept scheduling work. This is the HSF-476 stack: handleSnapshot → notify → forceStoreRerender.
• New createSnapshotBatcher (src/core/snapshot-batcher.ts). notify() is split into publishState() (updates getState()/getHandle() at once) and notifySubscribers(). For snapshots, the first one in a task notifies synchronously, and the rest of that task's snapshots fold into one setTimeout(0) notify. One batcher is shared by every subscription on a store, so a burst spread across many document subscriptions (one snapshot each) folds the same way as a burst on one collection. stop() drops the subscription's pending notify.
• Local writes (update/set/add/remove/sync) still notify synchronously, and a single snapshot behaves as before.

Risk: low. At most one task of delay before subscribers see the end of a burst. State reads always return the latest data.

How to verify

• pnpm typecheck && pnpm test. The new src/__tests__/snapshot-batching.test.ts fails on master: a burst of 100 snapshots, sent sync or through a microtask chain, notifies twice instead of 100 times, on one subscription or spread across 100 document subscriptions.
• I checked this against React 19.2 (jsdom, real createRoot, in hvakr). 60 rapid notifies plus a consumer useEffect(() => setState(value), [value]) throw #185 at notify 52. Routed through createSnapshotBatcher, there is no error and the final value renders. That was a scratch test and is not committed.
• A second scratch check in this repo (React 19.2, real hooks): 80 updates on one collection hook and one update each on 80 document hooks both throw #185 on master. With the per-subscription batcher only the collection case passed; with the store-wide batcher both pass.
• Follow-up in hvakr: release firestate and bump @hvakr/firestate in packages/app.

Link to Devin session: https://app.devin.ai/sessions/6640d72ea84e43d4aa4b46cf0c1f16bd
Open in Devin Desktop: https://app.devin.ai/desktop/session/6640d72ea84e43d4aa4b46cf0c1f16bd?variant=devin
Requested by: @andrewkrippner

---

<source media="(prefers-color-scheme: dark)" srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v3.svg"><source media="(prefers-color-scheme: light)" srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-light-v3.svg"><img alt="View with [code]smith" src="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v3-3x.png" width="162" height="28"> <a href="https://backend.blacksmith.sh/track/enable-autofix?expires=1795792068&installation_model_id=6841&pr_number=52&ref=codesmith_pr_footer&repository=flowcircuits%2Ffirestate&return_to=https%3A%2F%2Fgithub.com%2Fflowcircuits%2Ffirestate%2Fpull%2F52&signature=1a9993cf938b5931ec41570695985694c7d14c00ee8a0e40e06211ea2c41e217"><picture><source media="(prefers-color-scheme: dark)" srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-light.svg"><img alt="Autofix with [code]smith" src="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark-3x.png" width="190" height="28">
^{Need help on this PR? Tag @codesmith-bot with what you need.}

---

<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-devin-review-dark.svg?v=4"><img src="https://static.devin.ai/assets/gh-devin-review-light.svg?v=4" alt="Devin Review">

[devin-ai-integration]

I'll fix CI failures and address comments from users with write access. I'll skip comments containing "(aside)".

• Disable automatic comment, CI, and merge conflict monitoring

---

Original prompt from Andrew

HSF task HSF-476: Fix React #185 render loop on the Spaces canvas after Auto Takeoff

Full task body:

Finding from the Bug routine (CAR-2840). Not fixed: root cause not proven, and no local repro was possible (emulators unavailable on the agent box).
#``# Evidence (PostHog $exception, last 45 days)

• ~40 Minified React error #185 (Maximum update depth) events, ~15 distinct users, 2026-09-01 to 2026-10-09. Most are on /projects/<id>/building-model/spaces. A smaller set on /bod/space has a different stack (Table cell keydown, see closed #7338) and is out of scope here.
• Every Spaces-page event has the same stack, all library frames, no app frame:
  Firestore onSnapshot observer next → ``@hvakr/firestate collection `handleSnapshot` → `notify` → subscribers `forEach` → React `useSyncExternalStore` `forceStoreRerender` → `getRootForUpdatedFiber` throws `#`185.
• Strong correlation with Auto Takeoff: most events fire 10 s to 3 min after the same user's ai_space_takeoff_completed. One user also hit it while a takeoff was still in flight.
  #``# What the stack means
  React 19 throws #185 in getRootForUpdatedFiber when nestedUpdateCount > 50. The count rises on each commit that leaves Sync/Default work pending and resets only on a clean commit. The throw site is a fresh Firestore snapshot task, so the 50+ commits ran before it: something on the Spaces canvas keeps scheduling work after each commit while the spaces collection changes (server batch writes of suggested spaces, firestate autosave/rebase, suggestion reveal effects).
  #``# Ruled out (static read)
• SpaceSuggestor effects: each sets state only on a transition (0→N suggestions, reveal start/end).
• useMeasuredBBox and ScopeLabelView: both bail out on an unchanged size.
• SheetView processing shine: SMIL animation, no React state.
• useSpacesToDraw: reports each invalid space once.
  #``# Suggested next step
  Reproduce in the emulator: open the Spaces canvas, write 100+ suggested: true spaces to `proj... (1209 chars truncated...)

[devin-ai-integration]

✅ Devin resolved all 2 findings on 7a58a0a

Fixed by Devin (2)

• One subscriber blocks unrelated notifications
• Local updates trigger duplicate subscriber callbacks

View all findings in Devin Review