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
Commit 5ddcc74
Browse filesBrowse the repository at this point in the historyBrowse files
lumopolis: the ring of planned parts sits a slab's room out, so a part built from it rises in place even beside an outer district
Round 6b re-review of e935c1a:
- I1: the ring gap was 20, less than the 35 a one-building slab needs from
any district, so in a dense world the move was refused and the part fell
back to its layout slot. The ring now lies max(SLOT_CLEARANCE, the largest
district's radius + a one-building slab + MIN_SIBLING_GAP) past the
farthest district, and the moved district is centred on the wireframe, so
no district that stood before can refuse it. PLAN_RING_GAP is now
SLOT_CLEARANCE (the floor of that gap).
- M1: the ring is set from the districts as laid out, before any district
moves, so every wireframe of a compile shares one ring.
- M2: a moved district also keeps WIRE_ROOM (a one-building slab + the
sibling gap) from every wireframe still waiting, kept or new on the ring.
- M3: several moves in one compile are settled together: each is checked
against the others' new spots, and while one fails, the first failing one
(in matched order) stays put and the rest are checked again. Deterministic.
- Ring spots no longer hold -0, so they read the same after a JSON round trip.
Tests: an outer nine-building district at the wireframe's angle, a waiting
wireframe beside the target, a part added to the design as another is built,
and two ring parts built in one compile (both rise; with targets 20 apart the
first stays and the second rises).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
0 commit comments