The screenshot fingerprint and golden baselines stop being a tax on every
task branch:
- Task branches no longer commit docs/images/** or golden baselines. On a
branch the screenshot freshness stays a preflight warning; the review
prompt, REVIEWER.md and AUTHOR.md drop check-docs as a per-task gate.
- beta-derived.yml refreshes them on dev in one bot commit before the beta
candidate: canonical docs capture + docs:accept --reviewed, golden from
the golden-images artifact of a completed Validate on dev +
golden:accept --reviewed. A changed frame or scene is accepted only when
named in the inputs; undeclared differences refuse. Baseline commits carry
Release: and Baseline-Reviewed:; the subject is not a candidate subject.
- classify-changes: the Release: trailer on an issue/* branch no longer
switches on the heavy set. ci:full / ci:golden do: process-track emits
full=true, the review gate dispatches Validate with full=true and does not
accept a light proof.
Canon: PROCESS.md §3 п.13, §5.1, §8, §11.4; CONTRIBUTING.md.
Issue: #697
User-Visible: no
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
show/ship stop paying for diff mutants and for every move of dev:
- scripts/process-track.mjs resolves the track from the current labels and
the diff (show for unlabelled infra, ask for unlabelled product work) and
checks the mechanical ship limits; outside them the pipeline comments and
relabels track:ship -> track:show in the same round.
- Validate on the review material is light on show/ship: a completed push
run on the exact SHA is proof, a dispatch asks mutants=false. ask and the
ci:mutants label keep the mutant dispatch.
- show/ship skip the pre-review rebase when git merge-tree with dev is
clean; the candidate is rebased once at merge and still passes Validate
before the push to dev. The light merge waits for the push run of the
candidate and dispatches only when none appears.
- ship inside the limits merges after the light Validate without a model
review; the issue gets a machine marker hp:ship-merge.
- ship-review.yml + scripts/ship-review.mjs read the code of all ship
tasks of a beta range in one model session and publish
docs/reviews/SHIP-REVIEW-<tag>.md; both beta publication paths refuse a
range with ship tasks the document does not cover or that carries a High.
- show reviews judge correctness and AC; the spec review installs neither
npm ci nor Chromium, the show review installs Chromium only when the issue
names a smoke.
Canon: PROCESS.md §5, §5.1, §10.4, new §11.7; REVIEWER.md, AUTHOR.md and
AGENTS.md digests.
Issue: #696
User-Visible: no
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
The analysis of 85 closed tasks #600-#691 showed that the light track
cost as much as the full one (115 min and 12 events vs 102 and 13) and
that the owner had no label to choose the route. The owner accepted the
proposal on 2026-09-28.
- PROCESS §5 is the track table: track:ship (S1 -> S5, one line under
"## ТЗ", <= 30 src lines, batch review before the beta), track:show
(default, S2 -> S5, up to three AC, no spec review, 2 code cycles),
track:ask (full route). The owner's label beats the criteria, which
become a hint; any agent may raise a track, only the owner lowers it.
- §5.1: ci:full / ci:golden / ci:mutants order heavy checks on any track;
small and trivial read as track:show, no label as track:ask, an
infrastructure task as track:show.
- §2, §2.2, §2.4, §2.5, §4, §7.1, §7.2, §9, §11 follow; AUTHOR/REVIEWER
digests and AGENTS.md follow with the digest test and its mutants.
- task-packet.mjs reports the track via trackFromLabels(); the pipeline
reads track:show/track:ship for the cycle limit of 2 and lets an
explicit track:ask win. Pipeline behaviour by track is #696.
Issue: #695
User-Visible: no
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
The docs-screenshots workflow run 36430098138 (issue/689-daycycle-layers
before #691) decoded to the committed pixels in all 11 frames; its bytes
differ only by the oxipng repack. After the rebase onto #691,
`npm run docs:accept -- --identical` confirmed 11/11 pixel-identical
frames on the combined tree, so only the source fingerprint moves
(4955edd1 -> 62c5c18b); frames, their sha and the capture packer stay.
Release: v1.78.0-beta.9
Issue: #689
User-Visible: no
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
With the day/night background the plan went blurry after zooming from 100 %
(sharp when the page was opened at 800 %) and navigating a strongly zoomed
plan flashed the page white. Both came from #582's composition, not from the
wall hatch that #685 replaced:
- `.stage.daycycle.hp-safe-daycycle-outline .plan-svg` promoted the scene
with `will-change: transform`; Chromium freezes the raster scale of such a
layer, so the 100 % raster was shown stretched. The explicit layer #582
needs is now `will-change: opacity` (re-rasters at the current scale).
- The filtered outline had `overflow: visible` and a gesture exposed every
scene (#544) without bound, so the promoted layers grew with zoom squared
(CDP LayerTree, ~460 %: plan-svg 15.9x, outline 13.2x the stage; 39x after
navigating at 800 %). The full card clips its outline to its box and marks
it data-hp-live-overflow="clip" (never exposed); the live viewport bounds
every other exposure with an inline clip-path: inset(-25%) that leaves
with it, so idle DOM stays byte-identical (#531).
Owner-verified in Chrome 152 (built-in browser, DPR 2): sharp after 100 ->
800 %, no white flashes after reloading at 800 %.
Owner decision: #685's analytic gradient is reverted (13af1d5e), the single
<pattern> is back at every zoom; its close-up golden scenes stay and check
the pattern, its terminal-frame smoke checks the pattern.
Witnesses: demo/smoke_daycycle_zoom_layers.mjs (800 % x DPR 2: layers vs
stage, the hint, reload path, button/wheel/pinch); #582/#532 smokes now pin
the opacity hint; test/live-viewport.test.mjs (bounded exposure, clipped
scene); test/daycycle-layers.test.mjs (cascade). Four Node-guarded mutants.
Issue: #689
User-Visible: yes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
Все 11 кадров документации остались пиксельно идентичными; обновлён только fingerprint актуального дерева. Потолок initial View перецентрирован на 100 Б при факте 299713 Б, без изменения общего бюджета.
Release: v1.78.0-beta.8
Issue: #685
User-Visible: no
Все 188 golden-сцен и 11 кадров документации совпали пиксель-в-пиксель после переноса на актуальный dev; обновлены только паспорта текущего дерева.
Issue: #685
User-Visible: no
Release: v1.78.0-beta.8
Baseline-Reviewed-Local: sha256:c02067a968b83ee2ec2f110ab06fb3707491a79a98e61448f4b1ee8cc34c1a9a
WSL-приёмка обновляет 21 существующую масштабную сцену и добавляет четыре проверки 132/140%, DPR 1/2 с проёмами. Восемь кадров документации пересняты в той же доказанной среде; три неизменных кадра подтвердили совместимость.
Issue: #685
User-Visible: no
Release: v1.78.0-beta.8
Baseline-Reviewed-Local: sha256:414e55783bc8272d7b2ff209a94bd69e12de95f6759d5bc95975acb94e937f39
The Plan editor hid every device marker (display: none), so walls,
openings and stairs were placed against devices from memory. It now
shows them exactly as the Background editor does (#362): the same
markers as View, at the Background editor's effective opacity, fully
pointer-inert. The fade is filter: opacity(0.35) on each marker, not
on .devlayer, because room labels and room settings buttons share that
layer and stay opaque and interactive; filter multiplies the marker's
own opacity (.unavail -> 0.35 x 0.35), as the Background layer does.
Marker handlers already fail closed outside View/Devices.
demo/smoke_plan_device_landmarks.mjs proves AC1/AC2: same markers as
View, per-marker alpha equal to Background, core and capsule fall
through, a real click on a marker reaches the Walls tool and a room
settings button below, cursor and target unchanged, handlers inert, no
tabindex. Three mutants guard hiding, opacity override and the pointer
boundary. UX-MODES, DECOR-EDITOR, USER-GUIDE.ru mode table, changelogs.
Issue: #687
User-Visible: yes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
UX-MODES.md and TOUCH-SUPPORT.md said Reset (and, by omission, Esc) do
not finish an open Walls chain. The code finishes it on both
(houseplan-card.ts Escape handler, the tray's btn.reset), as does
WALL-THICKNESS.md §11 "Finishing a chain". Both documents now list the
same finish actions and point to §11; re-selecting Walls, pan, pinch,
a second pointer and pointer cancellation stay non-finish actions.
Issue: #684
User-Visible: no
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
Решение владельца 2026-09-28: «убрать рамку фокуса полностью». Навигационная
лестница View — фокусируемый `<g role="link" tabindex="0">` (#676), а своих
стилей фокуса у `.hp-stair` не было, поэтому браузер рисовал чёрно-белое
кольцо по прямоугольнику символа. Теперь `.hp-stair:focus` и
`:focus-visible` — `outline: none`: ни кольца браузера, ни замены. Лестница
остаётся в порядке Tab, Enter/Пробел по-прежнему переходят; hover и
выделение в редакторе не тронуты.
Свидетель — `demo/smoke_stairs.mjs`: снимок области лестницы с запасом на
кольцо без фокуса и с клавиатурным фокусом (`:focus-visible`) обязан совпасть
байт в байт; Enter на сфокусированной лестнице переводит на целевой этаж.
Без правила смок красный (`focusedStairPaintsNoFrame`). STAIRS.md и оба
changelog.
Issue: #686
User-Visible: yes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd
claude[bot]