Scope: the fourth chapter of the How Eve works track, rewritten in the Eve edition — sixteen slides moved out of three shared manuscripts into one, five chrome slides added, one new slide built on a capture that was already in the repository, and every binding re-reviewed. The brief records what the chapter is for and why each treatment was chosen; state.json holds the exact counts, durations and tool results.
What was checked#
- Build freshness.
build-eve-oshun.py --checkpasses on the delivered tree: 1,411 unique slides, 658 sources, 75 guides. The chapter is 22 slides in three sections of five, five and seven, running 30.3 minutes. - A real capture, not a new one.
assets/eve-product/work-exits.pngwas taken in the September fixture run at source8c27a8dand had never been used. Its bytes were checked againstprovenance.jsonbefore authoring, and no capture harness was run on this host — this machine's Chromium is 148 where every committed asset was made under 152, so a new capture would falsify the assets' recorded environment. - A legacy diagram retired.
eve-intent-to-artifactleftdiagrams.jsonfor a real graph diagram, leaving eighteen overrides. The test that used it as its text-wrapping fixture now usesv1-contracts-openapi, which still has a declared wrap width. - Assignments re-reviewed, not re-pointed. Five claims across three assignment files bind slides this chapter rewrote — the routing recipe, the add-a-tool recipe, the confirmation bridge, client actions and page control, and failure and fallback behaviour. Each was read against its new slide before the fingerprint moved, and each carries a dated note.
- Narration. All 22 tracks rendered and verified; the library-wide
--checkverifies 1,411 tracks (1,840.4 minutes). - Browser rendering.
eve-edition-review.mjs --guide eve-executionsweeps all 22 slides at 1440×900, 1280×720 with audio chrome, 390×844 and in print with no findings. Two slides needed fixing first: the unknown-outcome tree overflowed by 30px at 1280 until its node details shortened, and the capture slide overflowed by 56px until banner-shaped captures were given a smaller height budget on short screens. - PDF. All 22 pages pass the exact-text contract.
- Suites and lint. Python 456 passed; Node 443 of 445; Ruff clean; Prettier clean on everything changed. No JavaScript changed, so ESLint had nothing new to read.
What this chapter changed beyond itself#
- A CSS rule for every banner capture. A capture wider than three to one is short by construction, so spending the stack layout's full height budget on it only pushes the legend off a 720px screen. The three existing banner captures were swept again at every viewport with no findings.
- A rendering defect fixed in chapter four. On
eve-cap-delivery-board, three of the four numbered callouts sat on top of the column count they were pointing at, so the board read IN-PROGRESS, IN-REVIEW and VERIFIED with no number — on a slide whose whole point is which column the task is sitting in. Each badge moved to the blank right-hand end of its column header.
What was not checked#
- No human reader session and no listening review at normal speed.
- The narration was rendered on this host's CPU rather than the Apple GPU the rest of the library used; only the MP3 bytes differ.
- No new capture was taken, and the fixture behind the reused one was not re-run.
test_v2_rollback_sessionstill fails against its published baseline; re-baselining it is a decision for the user.