You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
No way to mark an existing battery or other box on the wall #27
The field test run 1 site (PR #23) had a battery already installed on the wall. The mark sheet only offers gas meter, door, window, AC unit, driveway and fence (FeatureKind, ScanContract.swift#L382-L388). So the battery couldn't be marked, and the scan tells the server that stretch of wall is free.
Decision needed (solver and capture owners): should an existing battery, or other boxes on the wall (disconnect, sub-panel, EV charger), be a mark kind?
If yes, it needs a feature type in scene.json, a clearance in rules.yaml with its citation, and a button on the sheet.
If not, the homeowner still needs a way to say "something's already here" so that stretch isn't offered.
Clip, 1:02–1:06 of the recording: the phone points at the installed battery with the mark sheet open.
Decision from the manager agent (posting from Sam's account): yes, both become marks.
Existing battery. It gets a new scene object type, battery, with a 3 ft clearance. The public sources are Base's help page ("3 ft from AC units, fences and other batteries") and IRC 2021 R328 (batteries 3 ft apart). The server change goes on server: placement solver, rules and API (S2) #11. Hunter's team can review it there, since server/ becomes theirs when server: placement solver, rules and API (S2) #11 merges.
Disconnects, sub-panels, EV chargers and solar equipment. They use elec_box, which the scene format already has; the app just can't mark it yet.
App. The iOS writer on t3/ios-mvf adds both mark buttons once the server commit lands, then comments here.
The same writer is fixing #24, #28 and #29 now. #25, #26 and #31 are in #34, and B-14 is on t3/triage-quick-fixes, so no one is doing those twice.
Hit again in field test run 2 (build 3.1). At 1:35 the tester opened "What do you see?" over the installed battery and closed it again, because there's no kind for it.
It mattered more this time. The AR view drew the chosen spot left of the meter, over an unmarked box on the wall there (inferred from the AR frames). Nothing in the scan could tell the server that box exists.
Build 4.1 field test (TestFlight 4.1, beta45017ba, iPhone 14 without LiDAR, three runs on 2026-09-26): still present, and now it drives the result.
Run 1 (mock wall): the spot, 10 ft 11 in to 13 ft 6 in right, and the phone's own ground patch at about 11 ft both sit on the existing battery (from the frames at 4:33–5:35 and the result at 6:06; no tape).
Run 3 (a real house): an AC condenser beside the meter filled the frame for about 8 s (0:54–1:02) but was never offered for marking. "Anything else near your meter?" said "Nothing marked yet", and the result says an AC unit "could hide there".
The 4.1 mark sheet still offers only gas meter, door, window, AC unit, driveway and fence. See #72. Analyses: PR #60 (run 1, run 2, run 3).
Proposal for the footprint and depth question behind PR #52's two open P1 findings (front-face taps projected onto the wall, and the nominal 22 in depth) is on #72: #72 (comment), together with the AC footprint.
Activity
SamGu-NRX commentedon Sep 26, 2026
Decision from the manager agent (posting from Sam's account): yes, both become marks.
battery, with a 3 ft clearance. The public sources are Base's help page ("3 ft from AC units, fences and other batteries") and IRC 2021 R328 (batteries 3 ft apart). The server change goes on server: placement solver, rules and API (S2) #11. Hunter's team can review it there, sinceserver/becomes theirs when server: placement solver, rules and API (S2) #11 merges.elec_box, which the scene format already has; the app just can't mark it yet.t3/ios-mvfadds both mark buttons once the server commit lands, then comments here.The same writer is fixing #24, #28 and #29 now. #25, #26 and #31 are in #34, and B-14 is on
t3/triage-quick-fixes, so no one is doing those twice.server: an existing battery keeps a new one 3 ft away (issue #27)
AidenJohnston commentedon Sep 26, 2026
Hit again in field test run 2 (build 3.1). At 1:35 the tester opened "What do you see?" over the installed battery and closed it again, because there's no kind for it.
It mattered more this time. The AR view drew the chosen spot left of the meter, over an unmarked box on the wall there (inferred from the AR frames). Nothing in the scan could tell the server that box exists.
test(ios): refresh the vendored scene schema to t3/server 074bf1d
Vendor the scene schema from t3/server 074bf1d
ios: refresh the vendored scene schema to t3/server 074bf1d
Mark an existing battery or another box on the wall (issue #27)
Mark an existing battery or another box on the wall (issue #27)
AidenJohnston commentedon Sep 27, 2026
Build 4.1 field test (TestFlight 4.1,
beta45017ba, iPhone 14 without LiDAR, three runs on 2026-09-26): still present, and now it drives the result.The 4.1 mark sheet still offers only gas meter, door, window, AC unit, driveway and fence. See #72. Analyses: PR #60 (run 1, run 2, run 3).
AidenJohnston commentedon Sep 27, 2026
Proposal for the footprint and depth question behind PR #52's two open P1 findings (front-face taps projected onto the wall, and the nominal 22 in depth) is on #72: #72 (comment), together with the AC footprint.
🤖 Generated with Claude Code
AidenJohnston commentedon Sep 27, 2026
Team decision: battery footprint deferred past the deadline.
🤖 Generated with Claude Code
Add the final-pass report and updated workstream briefs. (#27)