Lines drawn in the background (decor) editor showed notched 'teeth' where
two segments met at an angle. Decor <line> elements — both the saved render
and the live drawing preview — now carry stroke-linecap/linejoin=round, so
every line end reads as a circle of the stroke width and joints are smooth.
smoke_decor asserts the round attributes on the draft and the saved lines.
The cached config (LS_CFG) painted its first frames with _zoom=1 because
the saved per-space zoom was applied only by _restoreZoom()'s rAF after
the server config round-trip - the plan visibly jumped to the saved scale.
- setConfig now arms _zoom from LS_ZOOM for the resolved space before the
first view computation (view mode, no established view only);
- _restoreZoom applies the view synchronously when the stage is already
measured (space tabs, kiosk carousel, editor exit included); the rAF
path remains only for an unmeasured stage, where updated() already fits
with the correct zoom before the first paint.
New smoke_zoom_flash.mjs samples every rAF frame from the first possible
moment with a primed config cache, saved zoom 1.8 and a 350 ms server
delay: a visible stage with a default-scale viewBox fails the run
(22 such frames before the fix, 0 after).
Wall-midpoint handles for every room (finger-sized, pointer-captured, no
stage-pan conflict); dragging moves the wall along its normal with live
preview through the config snapshot, shared stretches drag the neighbour
(T-junctions insert vertices), openings on the wall travel along. Click a
room for the corner scale frame (uniform, about the opposite corner,
neighbours never dragged). Live badges: dragged+adjacent wall lengths and
m² per reshaped room. Grid snap, Esc cancels the drag, Ctrl+Z pops the
resize undo stack (one release = one step). Legacy rects are saved back
as polygons; three blob copies rebuilt.
HP-1543-01 (medium): exiting an editor AFTER switching floors inside it left
the editor working zoom on screen in view mode — the snapshot guard
(snap.space) dropped the restore and nothing else put the viewport back.
A space-mismatched exit now falls back to _restoreZoom() of the current
floor's saved view zoom (per-space store/LS_ZOOM were never polluted, the
view-only _saveZoom guard is untouched). Regression: crossFloor* asserts in
demo/smoke_zoom_out.mjs — red on v1.54.3, green now.
HP-1543-02 (low): a rapid off->on retrip before the current flash ended did
not restart the CSS animation: the senseflash class and animation-name never
changed, so the browser kept the old timeline and the second detection played
nothing once the first one-shot had finished (base opacity 0). Every
witnessed trip now bumps a generation counter; its parity alternates the
identical keyframes hp-sense / hp-sense-b via the .sf2 marker — a new
animation identity forces a fresh timeline per detection. flashTs and the
window timer re-arm as before; prefers-reduced-motion keeps the static ring
(retrip only extends the window). Regression: rapid* asserts in
demo/smoke_motion_sense.mjs — red on v1.54.3, green now.
Owner's contract (2026-08-01, вариант «б») replaces yesterday's 'sense' class
(29dae45, lived <1 day, stored nowhere — no migration):
- MOTION (device_class motion): a SHORT one-shot flash at the off→on
transition — 3 beats of the yellow ring (hp-sense 1.1s x3, ~3.3s), then
silence even while the entity still reports 'on': «движение = разовая
вспышка в момент обнаружения; cool-down не пульсирует». A new off→on trip
flashes again. Tracking lives in _senseRt (markerId → last state + flash
ts), stamped by _senseTick on the hass tick (the _vacTick pattern); one
setTimeout per entry repaints the card when the window closes so the
'senseflash' class is dropped without a state change (cleared in
disconnectedCallback). First sight already-'on' and unavailable→on do not
flash — not a witnessed detection.
- OCCUPANCY/PRESENCE while 'on': 'sensehold' — a STATIC yellow ring, no
animation, opacity 0.4 (the reduced-motion brightness): «присутствие =
статичное кольцо пока обитаемо».
- Unchanged: the yellow FILL stays reserved for «включено», alarm's red ring
wins on the shared ::after, ghosts draw nothing.
- prefers-reduced-motion: the flash becomes a static ring for the same ~3.3s
window (consistent with alarm); sensehold is static by design.
Smoke smoke_motion_sense.mjs rewritten to the new contract (13 asserts fail
on the pre-fix bundle): finite iteration count '3', flash gone after 3.6s
while state is still 'on' (the key assert), re-trip flashes again,
occupancy/presence static ring, neutral badge, hidden ghost inert.
shot_motion_sense.mjs now cycles off→on to arm the flash.
Owner's rule (2026-08-01): the yellow FILL is reserved for 'on' — a tripped
sensor keeps the neutral badge and gets a new 'sense' state class instead:
a gentle .dev.sense::after ring in the .dev.on yellow (--hp-on), 2.4s period,
max opacity 0.5 — the soft sibling of the red .dev.alarm siren. Declared
before .dev.alarm so red wins on the shared ::after if both ever apply;
ghosts don't pulse (hidden gate already strips the state class).
prefers-reduced-motion: static faint ring, same treatment as alarm.
Smoke smoke_motion_sense.mjs (fails on the pre-fix bundle): sense class on/off,
::after animation, neutral background (not --hp-on), occupancy/presence,
hidden ghost. shot_motion_sense.mjs captures the ring for review.
The exit-editor restore re-saved the view zoom from a rAF; on a slow
tablet the floor-tab click lands before that rAF (input runs first in
the frame), the snap.space guard skipped the fix-up save and the editor
wheel zoom (500%) stayed in _zoomBySpace/LS_ZOOM — the second floor
switch brought it back into view mode (owner's dacha report). Now
_saveZoom simply refuses to write while _mode is not 'view': the wheel
inside the editors keeps zooming but never touches the per-space view
store, so no fix-up is needed at all. The snapshot restore keeps
bringing the pre-editor viewport back. Smoke: smoke_zoom_out.mjs grows
the owner's exact scenario (both floors zoomed in view, editor 5.0,
same-tick switch after exit, two floor switches) — red before, green
now; the 6d16f69 asserts stay green.
Entering an editor snapshots the view-mode zoom+center (per space);
leaving back to view brings it back and re-saves it to LS_ZOOM, so wheel
zoom done inside the editors no longer leaks into the viewing zoom.
Editor-to-editor switches and the per-space view zoom keep working as
before. Smoke: smoke_zoom_out.mjs grows the 1.6 -> editor 2.5 -> 1.6
scenario (zoom, center, LS), the no-touch no-jump case and the
space-switch persistence guard.
Owner report (2026-08-01, screenshot): a device scaled up via marker.size
kept its value badge at the default size — an enlarged icon next to a tiny
'26.6°'. Same bug class as the v1.51.3 glyph fix (3456706): satellite
metrics were pinned to the base --icon-size, so --dev-scale grew the box
but nothing that belongs to it visually.
All .dev satellites now derive from var(--dev-size, var(--icon-size,
2.5cqw)) — which equals icon-size * dev-scale — keeping the exact same
coefficients, so at size=1 nothing changes pixel-wise:
- .valonly plate padding (0.16) and .valtext font (0.45)
- .tval/.hval plates: margin (0.1), radius (0.18), padding (0.14),
line-height (0.68) — font was already dev-size, the plate was not
- .lqi label: margin (0.05), font (0.38)
- .newdot: offsets (-0.12), diameter (0.34)
- .alarm ring inset (-0.35)
smoke_icon_scale extended: value badge and temp plate must double at
size=2 (getBoundingClientRect + computed font-size, 1.8-2.2 tolerance)
and keep the exact old defaults at size=1; verified failing on the
pre-fix build. space-card renders bare icons only, unaffected.
The v1.54.1 contract (first not-None value wins, zero is a value) covered
the source entity but not the card's fallback on the vacuum's own
selected_map: _vacMapId still used truthiness, so selected_map: 0 became
'default' on the frontend while trails.py resolve_map_id stored the run
under '0'. Calibration and server trails split across two keys and the
recorded run never rendered after reload.
The fallback is now the shared pure helper vacMapIdWithFallback (nullish
check), mirroring resolve_map_id. Cross-runtime regressions added for
selected_map = 0, '0' and '' on both sides; the frontend cases fail on the
old truthiness code.
Version 1.54.0 -> 1.54.1 in package.json, package-lock.json, manifest.json,
const.py and CARD_VERSION; rebuilt bundle in all three tracked copies
(dist/, demo/srv/assets/, custom_components/houseplan/frontend/).
Changelog entries (EN+RU) — one line per finding, HP-1540-01..06 — and
docs/STATUS.md bumped.
Suites on this exact tree: 161 frontend unit, 74 pure-backend, 74 browser
smokes — all green.
Audit HP-1540-02: the recorder chose the map id with an or-chain, so a
valid numeric map_index=0 fell through to selected_map (or 'default') and
the server stored the run under a key the renderer never looks up. The
choice is now resolve_map_id() — the explicit backend half of the contract
shared with vacMapIdFromAttrs in src/vacuum.ts: the first value that is
not None wins, zero and empty string included.
HP-1540-03: pairs was a plain source -> (marker, vacuum) dict, so the
second placement of the same robot (the documented two-floor case) evicted
the first and its server history silently stopped. A source now maps to a
list of pairs, every marker gets its own copy of the run, and the state
subscription set is deduplicated.
HP-1540-05: every config/set spawns async_refresh as a detached task; two
of them interleaving across the awaited config load could both subscribe,
overwriting one unsub handle — a callback leak until HA restart. Refresh
is serialized with an asyncio.Lock and teardown flags the recorder closed
first, so a refresh parked on its await can never resubscribe afterwards.
Regressions cover map_index 0/'0'/''/selected_map cross-checks, one
source feeding two floor markers across a map switch, pair-list refresh
with a deduplicated entity set, and two overlapping refreshes leaving
exactly one live subscription (zero after teardown). On the v1.54.0
recorder 11 of these tests fail.
At startup the vacuum entity reads unavailable; the recorder took that
for 'stopped' and ended the open run, so every HA restart mid-cleanup
rotated the trail into previous and began a fresh one (observed live:
a 21-point run became previous and restarted at 5). Unavailable and
unknown now mean 'no verdict'.
Caught live on the owner's X50 mid-cleanup with temporary logging: the
recorder saw every camera state change and rejected every one, because
server-side Tasshack keeps vacuum_position as a Point OBJECT — it only
becomes a dict when serialised to the frontend, which is why the card
adapter (and MCP inspection) always saw a dict and the stub test
faithfully reproduced the same wrong assumption. getattr fallback added,
regression test pinned, diagnostic logging removed (setup line kept at
INFO).
The plan shows the robot at work: the marker stays at its dock while a
puck drives the plan, calibration is one click or a drag-and-stretch
overlay, and the path is recorded server-side (current + previous run)
with never/cleaning/always display modes. Also: glow is the default
fill for new spaces, and the run-target search renders again.
Only TrailBook was covered; the HA-facing half — subscription callback,
attribute dialects, map-id resolution, run end on docking — had no test
at all. It does now, against a stubbed hass, which is also where the
missing behaviour showed up: recording started at the NEXT state change,
so an HA restart (or finishing calibration) mid-cleanup dropped the
opening seconds of the path. Sampling is factored out and runs once per
source on setup and on every refresh.
«Показывать путь робота» is a three-way choice now: never / while
cleaning (the default — the line hides the moment the run ends) /
always (the only mode that also shows the faded previous run).
trail_mode rides next to the legacy bool, which still maps in.
The last segment no longer pops in when the next telemetry point
arrives: a rAF sampler drags a tip line's endpoint to the puck's
animated centre every frame, so the path visually pours out strictly
from under the icon. The sampler runs only while pucks exist and stops
itself.
The integration now records the path itself (trails.py): it watches the
source entity's state changes, so recording needs no open card, has no
multi-tab write races, and every screen sees the same line — reloads
included, which retires the localStorage snapshot after one day of
life. Stored per marker: the current run plus exactly one previous
(owner call — users want cleaned-vs-uncleaned at a glance). The
previous run renders at 40% opacity; the current one still trims its
live tail so it never outruns the puck. Runs rotate on start or map
switch, points cap at 2000 with decimation, store writes debounce 10 s,
and houseplan_trail_updated pushes live cards. TrailBook is pure under
5 backend tests; the WS command degrades silently on older backends.
The self-recorded points now snapshot into localStorage per marker
(raw robot coords, so recalibration does not invalidate them). Restore
is gated: fresher than the linger window, same map, and never into a
run that started after the snapshot ended — the old trail must not
leak into a new cleanup. The smoke simulates the reload by wiping the
runtime map and asserts both the restore and the new-run discard.
The devlayer is pointer-events: none and every child opts back in; the
overlay never did, so every real click fell through to the plan while
the smoke's synthetic dispatch — which skips hit-testing — kept
passing. The overlay now opts in, the corner handles grew to finger
size, and the smoke performs REAL elementFromPoint hit-tests through
the shadow root so an untouchable overlay can never pass again.
Calibration is now a direct-manipulation overlay: the robot's rooms as
a dashed translucent ghost over the plan, dragged into place and
stretched by four corner handles (uniform scale about the opposite
corner). Quarter-turn and mirror buttons re-anchor about the ghost
centre; mirror defaults on because every robot map seen so far flips Y
versus the screen — measured on the owner's X50. Everything folds into
the same stored 6-number matrix, and legacy matrices reopen in the
panel with rotation snapped to a quarter. The park-the-robot-three-
times wizard is deleted outright: it was the most fragile part of the
feature (owner: «плохо работает»). fitMatrix/fitFromMatrix/initialFit/
reanchorFit are pure and unit-tested; the smoke drives the panel end to
end — drag, corner-stretch, rotate, save, puck on the new matrix.
Two owner reports. One: zoom, space switch or a tab return animated the
puck's left/top through the viewport change — it looked like the robot
driving across the whole plan. The view signature now forces the jump
class for that render, and a visibilitychange listener covers returning
to the browser tab. Two: a trail segment appeared the moment new
telemetry arrived, ahead of the still-gliding icon — the self-recorded
trail now lags exactly one point behind (the previous target is what
the puck has just reached), and an integration path is trimmed of its
live tail while moving.
The accent line vanished on same-hue fills. Blend modes (difference/
exclusion) all keep a blind luminance where the stroke disappears and
composite expensively on old kiosk WebViews, so the trail now uses the
cartographers' trick instead: a neutral dark halo under a light core —
one of the two always contrasts with whatever is underneath. Verified
over glow fill (dark rooms + light pools) in one screenshot.
ha-icon inside the puck lacked the .dev centering recipe (flex +
line-height: 0), so the glyph sat on its text baseline and floated
around the circle. The smoke now measures the glyph centre against the
puck centre to sub-pixel tolerance.
Owner's wording: «иконка похожа на иконку базы, только круглая и чуть
меньше» — same plate colors and shadow as a regular device badge
(var(--hp-bg)/--hp-line/--hp-txt), circle, 0.8 of the device size. The
smoke now compares the puck against a NEUTRAL badge computed-style for
plate parity — the robot's own base is yellow while cleaning and would
never match.
Checked against the real robot at the dacha: room centres arrive as
plain x/y next to the bbox, and the active-map name (selected_map,
'Первый этаж'/'Второй этаж') lives on the VACUUM entity, not on the
camera — without reading it both floors would silently share one
calibration matrix. Parser and card resolver adjusted; the captured
attribute shape is now a unit fixture.
The base marker never moves — it is the dock. While the robot cleans, a
round pulsing puck (no badge plate) drives the plan over an affine
transform solved from vacuum-map coordinates: auto-calibration matches
the robot's room list against plan rooms by name, and a three-point
wizard covers integrations without room data. The trail rides the
integration's own path when offered (it predates the card being opened)
and a self-recorded thinned buffer otherwise, lingering ten minutes
after docking. Adapters read the Map Extractor / Tasshack / Valetudo
attribute dialects through one tolerant parser. Display only — no
commands, per the owner's decision.
vacuum.ts is pure logic under 8 new unit tests; the marker schema grew
an optional vacuum block (56 backend tests); smoke_vacuum drives 19
browser asserts including the wizard end to end.
Owner call: 'Свет по источникам' is the mode that sells the card, so a new
space starts with it and the settings dialog offers it first. Deliberately
NOT changed: the fallback for an absent fill_mode stays 'none'
(spaceDisplayOf), so updating the card never repaints an existing plan
whose owner made no choice. smoke_space_settings re-pinned to the new
contract.
Reported by the owner minutes after v1.53.0: typing a name showed no
results. The results existed — .candlist is a scrollable box, and a
scrollable flex item inside the dialog body collapses happily: 26 matching
rows rendered inside a 1px strip. The binding dropdown never showed this
because it sits inside .droppanel, a block context.
flex: 0 0 auto + a min-height keeps it open. The smoke now MEASURES the
list and the first row instead of counting DOM nodes — counting is exactly
why it passed a build where nothing was visible.
The owner's spec shipped in dev yesterday, released as one:
- tap action 'Run automation/script/scene' with a searchable picker,
per-domain services, save/runtime guards for the target
- 'Ask for confirmation' checkbox guarding toggle and run alike
- covers/valves in the card-wide toggle, garage/door/gate excluded
Inventory: 148 / 52 / 43 / 72.
Owner's spec (2026-07-29), agreed points: one 'Run' action covering the
three runnable domains of HA (a script is the idiomatic 'action' — with
automations alone people would build trigger-less dummies); the confirm
checkbox guards BOTH toggle and run; covers and valves join the card-wide
toggle so curtains work natively.
- marker.tap_action gains 'run'; marker.tap_target (schema-bounded to
automation./script./scene. ids); marker.tap_confirm.
- the dialog: a searchable picker over the three domains (friendly name +
kind), save refuses a run action without a target, a vanished target gets
a warning hint; the checkbox shows for any actionable tap (explicit or
effective-default toggle).
- the tap: automation.trigger / script.turn_on / scene.turn_on, started/
error toasts; with confirm on — our own dialog (not window.confirm, it
must work on a wall tablet), Esc/backdrop/Cancel = no call. The guard
covers the controls-toggle path too.
- 'run' is explicit-only by construction: it needs a per-marker target, so
it can never arrive as a card-wide default.
- covers: the old test pinned 'garage stays shut' — that intent survives as
COVER_GUARDED_CLASSES (garage/door/gate stay out of the CARD-WIDE toggle;
an explicit per-device toggle remains the owner's conscious choice).
Locks/alarms stay forbidden everywhere, run included is not affected —
we do not inspect automation contents, same trust as HA's own Run button.
Tests: unit resolveTapAction/runServiceFor + cover guard, backend schema
parity picks 'run' automatically + tap_target bounds, smoke_tap_run with 11
assertions (picker, search, save guard, confirm cancel/ok, per-domain
services, missing target). smoke_tap_ctx: 4 options now.
Inventory: 148 / 52 / 43 / 72.
- HP-1521-01: the plan-mode assertion looked for ANY .dev.on and the lit
kettle satisfied it — a false positive hiding the very regression it
guards. It targets d_lamp now (kettle asserted separately), and the
mutation check proves it: reverting the v1.52.1 gate fails the smoke.
- HP-1521-02: the checklist entry and the _stateClass comment still said
'yellow in every fill mode'. Both now state the two-part contract: the
state predicate is the glow-pool condition; the renderer keeps the badge
only where the spot is not drawn.
- HP-1520-01: the glow layer is hidden in the plan editor, but the yellow
suppression still fired there — a lit lamp had NEITHER indicator. The
gate now equals the layer's visibility (disp.fill === 'glow' &&
!this._markup), so the badge returns exactly where the spot is absent.
- HP-1513-01: the static card ignored marker.size and marker.angle — the
same stored marker looked different on the two cards. It mirrors
--dev-scale and the icon rotation now; geometry only, no live dressing.
- HP-1520-02: TESTING/UX-MODES still demanded the removed RGB icon tint,
and the lightC comment described the old use. All three brought to the
v1.52.0 contract.
smoke_light_badges grew the editor-mode vectors; new
smoke_size_angle_parity asserts the x3 ratio inside each card (absolute px
are incomparable across containers) and rotation on both. Inventory:
147 / 51 / 43 / 71.
Owner's rule, agreed 2026-07-29 after a field report (a lamp turned off by
tap looked different from one turned off by the wall switch):
- a lamp's colour lives ONLY in its glow. The v1.27 RGB tint of the icon,
border and shadow is deleted — that tint was the fork: with colour data
the lamp rendered dark-with-coloured-icon, without it plain yellow, and
the same lamp crossed the fork depending on how it was switched.
- in glow fill the indicator IS the spot: a source's badge stays standard,
lit or not (litLightEntity — the exact condition that casts the spot —
gates the suppression, so a lit socket keeps its yellow even in glow).
- in every other fill a lit source is plain yellow, like a heating TRV.
- icon morphing stays everywhere; the ripple colour still falls back to the
light colour (both explicitly confirmed by the owner).
smoke_light_badges covers the whole table (8 assertions); smoke_rgb_alarm
re-asserted: no rgb class, lit lamp yellow, ripple fallback keeps the
colour. README colour language updated. Inventory: 147 / 51 / 43 / 70.
- HP-1511-01: defaultPositions ran over different rosters — the full card
reserves grid cells for hidden devices, the static card compacted them
away, so an undragged marker sat in different spots on the two cards. The
static card feeds spaceDevs (hidden included) to the shared grid and
renders devs (visible) — exactly the split HP-1510-01 introduced for LQI.
- HP-1511-02: a hidden ripple-display marker rendered as an icon-less
inactive pulse. A ghost drops the display dressing entirely: ripple
presentation off, noicon off, base icon on, whatever marker.display says.
smoke_hidden_flag: the weak 'has icon OR noicon' assertion is gone — every
ghost must carry a base icon; new autoGridParity vector (vb-coordinate
comparison, the cards render in different view systems) and a ripple-ghost
vector. The demo stub got a connection.subscribeEvents so the static card's
module-level config cache can be invalidated between in-test cards.
- HP-1510-01: the static card's visibility filter had quietly become its
aggregation filter — the same room showed different Zigbee health on the
two cards. Two lists now: aggregation (room LQI, temp) sees every device
of the space including hidden ones, rendering sees visible only. Light
fill keeps excluding hidden through areaLights itself, so the contract
stays exactly as agreed: hidden counts toward signal, casts no light.
- HP-1510-02: the ghost suppressed state colors but still painted value
text, temperature, humidity, the LQI badge and the state-morphed icon.
All live numbers are gated on d.hidden now — a ghost is the base icon and
the name, nothing else.
smoke_hidden_flag grew both audit vectors: the 42 kW value-display ghost
renders no numbers, and a room whose only Zigbee devices are hidden paints
the identical lqi fill on the full and the static card.
The dev batch since v1.50.4, released as one:
- hiding is a per-device checkbox seeded once from the old filter
(docs/FILTERING.md); blue ghosts under a local 'Show hidden' toggle
- yellow = doing its main job now; TRVs glow by hvac_action, service
switches can no longer become a device's primary
- pinch/pan gestures in every editor on touch
- the room settings button: visual centre (inscribed circle + centroid
pull), icon-derived size, zooms with the plan; metrics visible in the
plan editor
Inventory: 147 frontend / 51 pure / 43 harness / 68 smokes.
The full card's _buildModel() was a hand-copied twin of spaceModels(), and
the twin missed the legacy-store fallbacks v1.50.3 gave the shared builder —
the same broken store rendered recovered in the static card and as
viewBox='0 0 0 0' with negative-width rects in the main one. The divergence
of the duplicates IS the bug, so the duplicate is gone: the full card calls
spaceModels() and only swaps the raw plan url back in (its signing flow must
not bake a signed url into a memoized model — 2026-07-27).
New smoke_legacy_geometry runs the audit's exact vector (zero viewport +
negative rect) through both models and both DOM trees and asserts parity:
full-canvas fallback, normalised rectangle, no negative SVG attributes.
Inventory: 140 / 51 / 43 / 65.
The ±4 bound from v1.50.2 measured view_box[2:4] and room w/h with the same
ruler as coordinates, so zero and negative sizes still passed the schema —
and viewBox='0 0 0 0' draws nothing on every client, with the static card
computing aspect-ratio: 0 / 0 on top. _EXTENT now requires strictly positive
sizes with a floor of one thousandth of the canvas (1 render unit — far
below any real room, keeps the maths finite); coordinates stay allowed to be
negative, a crop origin legitimately sits past the edge.
Defensive layer for stores that already hold a broken viewport: spaceModels
falls back to the whole canvas — both cards render from that model, so both
get the fallback — and a legacy rectangle with a negative size reads as the
same rectangle drawn from the other corner.
Also: the room settings button is the bottom row of the room card, and the
room name renders in the same spot in view and plan modes (owner's request,
committed earlier on dev).
- HP-1501-01: v1.50.1 bounded layout positions and left room rectangles,
polygon vertices, view_box and opening coordinates on bare _finite — the
same absurd-magnitude failure, one schema over. _GEOM (±4) covers them all
now, opening angles get ±360. And because a store may already hold such a
vertex from before the door existed, contentBounds applies its canvas
envelope to room geometry exactly as it does to device positions: the
point renders where it is, the frame ignores it, a space of nothing but
absurd points falls back to the whole canvas.
- HP-1501-02: a repair matching zero positions answered ok/moved:0 and
replaced the one-deep backup with an empty one — a typo right after
repairing the wrong space destroyed the promised way back. Empty match is
nothing_to_repair now: no write, no revision bump, backup intact.
Old test fixtures carried view_box [0,0,100,100] from the render-unit days;
they now use the normalised box the product actually stores.
- HP-1500-02: the stage budget was the absolute document coordinate, so any
tall dashboard content before the card was billed as header and the stage
collapsed to 0px. Measure our own chrome relative to the card plus a
bounded (<=120px) allowance for what the viewport keeps above us; re-measure
on window resize, remove the listener in disconnectedCallback.
- HP-1500-03, both layers: contentBounds opens a near-zero axis (< ~an icon)
up to a 200-unit floor and ignores extra points outside a canvas envelope
(-25%..125%) for FRAMING purposes only; the server bounds layout coordinates
to +-4 — any finite float used to pass, and one 1e100 hid the plan from
every viewer. A thin real room keeps its tight frame; the gate sensor past
the edge still stretches it.
- HP-1500-01: no automatic double-transform — a correct layout and a stranded
one are indistinguishable, and guessing wrong corrupts good data. Explicit
admin command houseplan/geometry/repair: dry_run previews, the backup rides
the same store write, undo restores, and routine layout writes now preserve
unrelated store keys instead of eating the backup.
Tests: contentBounds guards (unit), layout coordinate bounds + repair
lifecycle (harness), card-below-content smoke. Inventory: 139 / 49 / 42 / 64.
Owner's batch (committed to dev earlier today, released here):
- devices count as content for the default zoom;
- the editor no longer shifts the plan — the stage measures its own top
instead of assuming 118px of header;
- zoom goes out to 0.4x, centred.
From the review:
- HP-1490-01: the square-canvas migration wrote two stores in sequence, and
the first write deleted the aspects the second needed — a crash between
them stranded the layout in the old coordinates with nothing able to
finish it. The intent {space: old aspect} is durable now: saved to the
layout store before anything moves, cleared by the same write that stores
the migrated layout, each half idempotent behind its own trigger. The
update event fires only after both halves are on disk. Proven at the exact
crash boundary by a harness test that fails the layout write once.
- HP-1490-02: check_quota and the file write were two executor jobs with
nothing between them, so N parallel uploads all measured the store before
any of them wrote. One job under a dedicated upload_lock now — narrower
than write_lock on purpose, a directory scan must not stall config saves.
A failed write reserves nothing.
- HP-1490-03: the content frame fed pan, zoom, clamp AND pointer maths, so
the editors were boxed into yesterday's drawing. Edit modes measure from
the full square; mode switches refit rather than carry a view clamped
against the wrong base.
- HP-1490-04: Save could outrun the proportions read and ship the previous
file's ratio. Picking a plan clears it immediately; Save awaits the
bounded read and stores 'unknown' over a lie.
- §5: package-lock version synced, duplicated comment removed.
New: smoke_audit_1490.mjs, migration crash-recovery pure + harness tests,
parallel-quota harness test. Inventory: 138 unit / 49 pure / 40 harness / 64
smokes.
Codex