# M10-11 Status

status: done
task: NLA track/strip/action/time mapping read-only exact evaluation
updated: 2026-08-16 America/New_York

## Scope

A Blender 5.2 desktop-authored fixture contains one object, one NLA Track and two
Action Clip strips. The first maps Action frames 1..11 to scene frames 20..40 with
scale 2. The second maps the same Action range to frames 45..65 with reverse playback
and repeat 2. Both use `NOTHING` extrapolation so samples outside their ranges expose
boundary behavior rather than hold behavior.

The Main reader must preserve Track, Strip, Action and time-mapping fields. Blender's
native Depsgraph evaluates 12 boundary/interior/cycle frames; Node WASM and the
production Chromium Worker compare every world matrix with the desktop golden. The
test asserts that evaluation does not change NLA or Action JSON. It does not call
`setNLAStack` and does not claim the M10-12 operator transaction.

## Evidence

- Fixture: `tests/files/web/nla_time_mapping_scene.blend`, SHA-256
  `b32f783debe213a1345ec4528e5ed27f0f4de00d7896b626df39756289119855`.
- Golden: `tests/golden/M10-11/nla-evaluation.json`; Blender 5.2.0 LTS, 1 Track,
  2 Clip strips and 12 frames.
- `WEB_TEST_PORT=5527 npm --prefix web run test:nla-evaluation-golden`: Node/WASM
  checker and Chromium 1/1 passed; maximum matrix error 0 and read-only identity passed.
- `npm --prefix web run typecheck`, `npm --prefix web run lint`, and
  `npm --prefix web run check:local-deps`: passed.

## Implementation Hashes

- Fixture generator: `a941ecb850ddd165ede117f6411615c6daed4d6bdaffaeac8228a308b7fbe745`.
- Golden generator: `941a63cc9edbaa0d8dd25025872e3fc44bf64df54d5a9232ad4128e6ac0cd5c5`.
- Node/WASM checker: `2ce9db0fcbceaf30e398e8feb5c002a60395ac34e61382d9414e980e51e6fdd5`.
- Golden: `2d148ac0cbc2ac9281d430da20dba8373ad1bb0de6ec5fa578adf36296455fbe`.
- Chromium spec: `c12f105d206a03dae5b73b076cc3b0a758d597cea7532b6720f4653568df0245`.

## Rollback

Remove the fixture/golden generators, fixture, M10-11 golden/checker/Chromium spec,
package command and this status file together. Restore the M10 count to 10/15. Keep
the existing N-014 bounded Main writer and reverse/repeat regression unchanged.