test(studio): characterization suites for resize commit and razor history

[ukimsanov]

What

two test-only suites pinning CURRENT behavior before the NLE swap: anchoredResizeReleaseShift.test.ts (manual-offset resize release commits) and useRazorSplit.history.test.tsx (razor split undo/redo history).

Why

regression tripwires — the later glue-swap PRs must keep these green.

How

test files only; they import existing main modules unchanged and pass against them as-is.

Test plan

bunx vitest run on both suites; fallow audit clean.

---

Stack position 10/25 — studio NLE overhaul (CapCut-parity timeline + canvas). Graphite manages bases; merge from #2192 upward. Temporary TEMP(studio-dnd) fallow entries (unwired-component windows) are all removed by #2213 (app-shell swap), whose tree is byte-identical to the fully-verified integration branch.

[ukimsanov]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• feat(studio): assets/blocks panel behaviors + preview helpers #2216
• feat(studio): timeline visual refresh #2215
• feat(studio): clip thumbnail modules #2214
• feat(studio): app shell swap — mount EditorShell, retire the legacy engine #2213
• feat(studio): timeline glue swap — Timeline/TimelineCanvas onto the NLE engine #2212
• feat(studio): canvas glue swap — overlay, gestures, selection to final form #2211
• feat(studio): sidebar asset card (unwired) #2210
• feat(studio): NLE shell assembly (unwired) #2209
• feat(studio): canvas chrome components (unwired) #2208
• feat(studio): timeline interaction hooks and lanes component (unwired) #2207
• feat(studio): clip drag engine — preview and commit (unwired) #2206
• feat(studio): glue API coexistence layer for the NLE swap #2205
• feat(studio): NLE shell components — context provider, panes, overlays (unwired) #2204
• feat(studio): arrow-key nudge step math (unwired) #2203
• feat(studio): canvas context menu and z-order actions (unwired) #2202
• test(studio): characterization suites for resize commit and razor history #2201 👈 (View in Graphite)
• feat(studio): asset click policy and canvas nudge gate #2200
• feat(studio): timeline lane-zone model #2199
• feat(studio): timeline z-stacking sync model #2198
• feat(studio): multi-clip drag preview math #2197
• feat(studio): timeline magnetic snapping #2196
• feat(studio): timeline collision and placement model #2195
• feat(studio-server): files route extensions #2194
• feat(studio): seek-restore contract for player reloads #2193
• feat(studio): timeline leaf helpers — audio inspector, zoom math, UI prefs #2192
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[miga-heygen]

Reviewed as part of the 25-PR NLE overhaul stack. Full stack review on #2205 (the keystone). No blockers on this PR. — Miga

[vanceingalls]

🟠 Request changes — useRazorSplit.history.test.tsx is genuine characterization and lands 🟢; anchoredResizeReleaseShift.test.ts is 2/3 aspirational math theater that cannot regress on real code changes.

useRazorSplit.history.test.tsx — 🟢 legit characterization.

Both branches (plain HTML + GSAP rewrite) mount the real useRazorSplit hook, exercise handleRazorSplit under act, and assert (i) the split reached disk via the mocked fetch layer that faithfully mirrors /api/projects/:id/file-mutations/split-element/ and /api/projects/:id/gsap-mutations/, (ii) snapshot().undoLabel === "Split timeline clip", (iii) an undo via undoViaDisk restores the exact pre-split bytes. If the production hook drops the history entry, uses a different label, or fails to route through the persistent edit history store, the test fails. Faithful.

Two small callouts, non-blocking:

• The GSAP-branch fixture (SPLIT_GSAP, line 271-275) writes tl.set("#clip1-split",{x:0},2) — a static synthetic timeline. The server-mirror mock at line 297-306 hardcodes the response to SPLIT_GSAP regardless of what the hook actually asks for. This pins the "endpoint gets called + response gets applied" contract, not "the server produces this exact rewrite given this input" (which is a studio-server concern, correctly out of scope here).
• expected: finalContent (line 356) uses the fixture as the equality target — meaning the test pins that studio's fetch pipeline round-trips the server response to disk verbatim. Legit contract for a client-side hook test.

anchoredResizeReleaseShift.test.ts — 🟠 aspirational, does not pin production behavior.

Tests 1 (line 97 "the per-frame center anchor converges") and 2 (line 139 "the center stays fixed for every corner at any rotation") both simulate an N-frame anchor loop using a test-file-local helper accumulateResizeAnchor (line 50-60). That helper doesn't wrap or import any production symbol — it re-implements the fix's math inline. The tests then assert that this local math converges.

I grepped the repo for the fix's key state (lastResizeAnchor, accumulateResizeAnchor) at this PR's head SHA — zero hits in production code. The docstring at lines 39-41 says "Fix (useDomEditOverlayGestures pointermove, resize branch, fa4f39168)" — that SHA doesn't resolve in the repo history, and useDomEditOverlayGestures.ts at this SHA has no resize-branch anchor accumulator (grep shows only one line 384 unrelated).

Consequence: these two tests will pass whether the production resize gesture converges, oscillates, or is deleted entirely — because they never call into it. As a regression tripwire for the "later glue-swap PRs must keep these green" claim in the PR body, tests 1 and 2 are false safety. Two possible fixes:

1. Preferred. Import the actual production accumulator from useDomEditOverlayGestures and drive the loop through it — even if that means factoring the per-frame anchor step out to a named computeNextResizeAnchor(prev, fixedStart, centerNow) helper that the gesture hook calls. That's the "leaf helper" pattern this stack uses everywhere else (see #2192).
2. If the production fix hasn't landed yet in this stack: delete tests 1 and 2 from this PR and re-add them in the same PR that lands lastResizeAnchor on the hook, so they're tied to real code from the start. Landing aspirational tests ahead of the production fix creates the exact false-safety window the PR body's "regression tripwires" framing is designed to prevent.

Test 3 (line 182 "net translate after persist+reload equals the committed anchor offset (non-GSAP)") does import real production symbols (applyStudioBoxSize, applyStudioPathOffset, readStudioBoxSize, reapplyPositionEditsAfterSeek, createManualOffsetDragMember, applyManualOffsetDragCommit, buildBoxSizePatches, buildPathOffsetPatches), asserts the committed anchor offset survives the persist→reload chain end-to-end, and would fail if any of those functions drifts. Keep test 3 as-is.

Ask. Either (a) rewire tests 1+2 to drive the loop through the production accumulator this PR body says exists, or (b) drop tests 1+2 and land them alongside the fix in a later PR. Either is fine — the current state ships fabric that looks like characterization but isn't.

R1 by Via

[ukimsanov]

Fixed — you were right that tests 1–2 proved nothing about production: the accumulate math only lands with the gesture code in #2211, so a test-local helper here was false safety. Taken your option composite: tests 1–2 are removed from this PR (test 3, which imports the real persist symbols, stays), and #2211 now extracts the per-frame anchor step into a named production helper (computeNextResizeAnchor in useDomEditOverlayGestures.ts, same math, one call site) with the convergence tests rewired to drive it directly — they now fail if the production loop oscillates. Also fixed the stale SHA reference in the docstring.