Skip to content
Source
Sign inSign up
TanStack/db

fix(db): compose optimistic updates from changes instead of whole-row snapshots #1701

Closedmarbemac(GitHub profile, opens in a new tab)wants to merge 1 commit intomain fromrefs/pull/1701/headopened

🎯 Changes

Small fix to how the optimistic overlay composes rows.

Today an optimistic insert/update stores mutation.modified — the whole row as it looked at mutate() time — into optimisticUpserts. Since modified is built from the visible row (Object.assign({}, visibleRow, payload) in mutations.ts), whichever transaction sorts last dictates every field, including ones it never touched.

The clearest symptom: rolling back one of several in-flight transactions doesn't actually remove its change.

It doesn't change when committed sync gets applied. Sync still stays queued while a transaction is persisting, exactly as today. That's the half of #1630 that turned out to be unsound without temp/server key mapping, and it's not needed for any of this.

So this is only the projection half — same base as before, just composed per-field instead of replaced wholesale.

Related to RFC #1625.

// three transactions in flight on the same row
t1.mutate(() => collection.update(1, (d) => { d.a = `a1` }))
t2.mutate(() => collection.update(1, (d) => { d.b = `b2` }))
t3.mutate(() => collection.update(1, (d) => { d.c = `c3` }))
t1.commit(); t2.commit(); t3.commit()

t2.rollback()
// expected: { a: 'a1', b: 'b0', c: 'c3' }
// actual:   { a: 'a1', b: 'b2', c: 'c3' }   <- b2 is back

t3's snapshot was taken when b was already b2, so once t2 is gone t3 puts it right back. I don't think there's a reading of that which is correct.

The fix: update mutations now compose mutation.changes over the value they were layered on, so a mutation only owns the fields it actually changed. Inserts keep using modified, since an insert's changes deliberately omit schema defaults (there's a comment in mutations.ts saying as much).

✅ Checklist

  • I have tested this code locally with pnpm test.

🚀 Release Impact

  • This change affects published code, and I have generated a changeset.
  • This change is docs/CI/dev-only (no release).

Summary by CodeRabbit

  • Bug Fixes
    • Fixed concurrent optimistic updates so each transaction changes only the fields it modified.
    • Prevented rollbacks from overwriting unrelated updates or leaving stale values visible.
    • Preserved optimistic insert defaults when subsequent updates are applied.
    • Prevented redundant change notifications when an optimistic overlay produces no effective change.
    1. Review Change Stack

      No actionable comments were generated in the recent review. 🎉

      ℹ️ Recent review info
      ⚙️ Run configuration

      Configuration used: defaults

      Review profile: CHILL

      Plan: Pro Plus

      Run ID: e26dd938-b40d-4dee-ada3-fc6733579ea1

      📥 Commits

      Reviewing files that changed from the base of the PR and between 67c840fe7a45d818de768fa790ebec2f27654d04 and a18618772e95ba80ef6ff484c6ffab11ff2e6285.

      📒 Files selected for processing (3)
      • .changeset/soft-donkeys-shave.md
      • packages/db/src/collection/state.ts
      • packages/db/tests/optimistic-composition.test.ts

      📝 Walkthrough

      Walkthrough

      Optimistic updates now merge only changed fields over the current row, preserving unrelated fields across concurrent transactions, sync commits, and rollbacks. Tests cover composition, defaults, synchronization, rollback isolation, and redundant event suppression.

      Changes

      Optimistic composition

      Layer / File(s)Summary
      Compose optimistic upserts
      packages/db/src/collection/state.ts, .changeset/soft-donkeys-shave.md
      CollectionStateManager resolves update mutations by merging changes with the current base row, caches valid compositions, and uses the resolver during optimistic recomputation and sync commits.
      Validate overlay behavior
      packages/db/tests/optimistic-composition.test.ts
      Tests cover overlapping updates, unrelated sync changes, rollback isolation, retained schema defaults, and redundant event suppression.

      Estimated code review effort: 3 (Moderate) | ~20 minutes

      🚥 Pre-merge checks | ✅ 5
      ✅ Passed checks (5 passed)
      Check nameStatusExplanation
      Title check✅ PassedThe title clearly summarizes the main change: composing optimistic updates from field-level changes instead of whole-row snapshots.
      Description check✅ PassedThe description matches the template and includes changes, checklist, and release impact with testing and a changeset noted.
      Docstring Coverage✅ PassedNo functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
      Linked Issues check✅ PassedCheck skipped because no linked issues were found for this pull request.
      Out of Scope Changes check✅ PassedCheck skipped because no linked issues were found for this pull request.
      ✨ Finishing Touches
      🧪 Generate unit tests (beta)
      • Create PR with unit tests

      Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

      ❤️ Share
      Comment `@coderabbitai help` to get the list of available commands.
    2. I checked collection/state.ts and confirmed both places that seed optimisticUpserts for an update mutation just did mutation.modified as TOutput, a full snapshot, not a diff. That's the exact bug: two transactions mutating different fields on the same row would fight over which snapshot wins the overlay, and rolling back one would blow away fields a sibling transaction owns, since there's no per-field ownership tracked anywhere. resolveOptimisticUpsert composing {...base, ...mutation.changes} fixes this cleanly. The caching by the changes object reference plus a base === cached.base check is a nice detail, it keeps object identity stable for an unchanged overlay so downstream consumers relying on reference equality (e.g. to suppress redundant update events) don't get spurious churn, which the last test in the suite explicitly checks. I'd double check the WeakMap doesn't leak in long-lived collections with high mutation churn, but since it's keyed by the changes object (per-mutation and short-lived), that seems fine.

    3. Closing as superseded by #1807. The current optimistic-update contract intentionally uses captured whole-row snapshots, with oracle coverage for that behavior; applying this PR now would create hybrid rows that neither the mutation nor the server produced. RFC #1625 remains the place for any future redesign of reconciliation semantics. Thanks for working through this.

Sign in to join this conversation.