[vivekjyani]

Context

DocumentEditor.makeBulkChange(from:) consolidates a batch of per-row changelog events into a single v:2 bulk event (bulkRowUpdate / bulkRowDelete). Previously it required changes.count > 1 and returned nil for a single change, forcing callers to special-case the count-of-1 path. A single change needs no consolidation, so returning nil was surprising.

What changed

• Sources/JoyfillUI/ViewModels/DocumentEditor+ChangeHandler.swift — makeBulkChange(from:): empty array still returns nil; a single change now returns that original event unchanged (v:1, original target); 2+ changes consolidate into a v:2 bulk event exactly as before.
• JoyfillTests/.../DocumentEditor+ChangeHandlerTests.swift — the two single-change tests now assert the original event passes through (target unchanged, v == 1) instead of asserting nil.

Decisions

• Pass the original Change through as-is rather than wrapping one change in a bulk envelope — the consumer gets the plain event it already knows how to handle, and no needless bulk* target is emitted for a single row.
• Empty-input behavior is intentionally left as nil (nothing to send).

Screenshot / video

N/A — logic-only change, no UI.

Public API

makeBulkChange(from:) is public; the signature is unchanged. Only the single-change return value changes (nil → original event). Minor doc note: the method's doc comment still reads "Returns nil unless it's a multi-row…"; worth a one-line update when touched next.

Tests

Covered. Single-change update/delete passthrough is asserted; empty-array nil, multi-row consolidation, and all existing guards remain green (132/132 in DocumentEditorChangeHandlerTests).

Reviewer notes

Small, self-contained follow-up to the already-merged makeBulkChange feature. No behavioral change for the empty or 2+ paths.

[vivekjyani]

Reviewed the single-change passthrough change. One contract question and two small cleanups below.

[VishnuBishnoi]

Reviewed at eaab942 against 616d3d6.

No blocking findings. The target guard now runs before singleton passthrough, so only rowUpdate/rowDelete are returned unchanged. The full original Change is preserved, and the 2+ item validation/consolidation paths are unchanged. The documentation matches this behavior.

Exact-head build and unit CI passed: 2,334 tests, zero failures. UI tests were skipped; this change is confined to the event helper. No local tests were run.

Non-blocking test improvement: assert full singleton dictionary equality, including nested routing and extra metadata, rather than only target/version.