beezm
0e39538604
All three gate polish items — they interleave in the same files, so they land as one buildable commit rather than three that do not compile independently. 1. TRIBUTARY WATER. Tributaries carved in task 22/23 but stayed dry: they were never registered with the water stage. They now are (as CarvedRiver entries named 'trib', kept OUT of the per-river console table so it stays the 6 mains). RiverTribWaterMinFlow (40k px of along-course flow) sets where water starts; RiverTribTaperPx (120 px) makes the wet->dry transition a FADE, not a wall — across the taper the wet strip narrows (0.25x -> 1x half-width) AND its surface drops toward the bed, so a stream head thins out and vanishes. Along-course flow is modelled quadratically from headwater trickle to full drainage at the mouth (the plan records drainage per river, not per sample). RiverTribWaterMinFlow=0 waters them end to end — the documented fallback. 2. LAKE-ENDER TARGETING. The join routed to the nearest classify-water CELL, which a 322-px puddle satisfies; it stopped ~80 px short of the 197k-px lagoon beside it. It now routes to a SIGNIFICANT water mask (bodies of at least RiverLakeMinTargetPx = 20k cells, built from the water-body table), falling back to any classify water only if no significant body is reachable, so a seed with genuinely small ponds still connects. Proven by adjacency: task 23 had 63 river cells touching the puddle and 0 touching the lagoon; task 24 has 49 touching the lagoon and 0 touching the puddle. 3. FINER STEPS + DEEPER BEDS. RiverStepDropM 2.0 -> 0.6 and RiverWaterDepthM 1.2 -> 2.2, RiverDepthScale 1.0 -> 1.5. Reaches 298 -> 1034; the level gap between adjacent reaches drops from median 0.571 m / p95 2.08 m to median 0.124 m / p95 0.74 m — the pond-staircase reads as a graded descent. Reaches stay trivial bodies (median 44 px). Guards all held: flood guard 0 newly-below-sea and 0 below-sea cells modified, BIOME oracle md5-identical to the task-22 baseline (output-only unchanged), island top 457.65 m exact, crater core excluded, lowest carved cell exactly sea+margin (37.85 m). Cumulative max cut 25.98 m with 339 cells >20 m — identical to task 23, so deepening the WATER did not deepen the worst cuts. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| Scenes | ||
| Scripts | ||
| README.md | ||
/Tools
Offline developer utilities. Nothing here ships. These generate the source data that /Server
and /Client later consume, which keeps world-generation logic out of the shipped client.
The 2D world generator
Script: Tools/Scripts/MapGenerator.cs · Scenes: Tools/Scenes/MapPreview.tscn and
Scenes/MapCaptureTool.tscn
Builds the topographical and logistical foundation of the island before any 3D exists.
- Resolution is config-driven, from
ServerConfig.json'sMapProfile:4K/6K/8K/10K→ 4096 / 6144 / 8192 / 10240. Shipped default is 8K. 1 pixel = 1 in-game metre = 1 voxel footprint. (TheMapSizefield initialiser in the script says 4096, but it is overwritten from config before use.) - Topography: FastNoiseLite plus distance-falloff maths for a guaranteed mainland island, a mountain spine, dynamic temperature-driven sea level, and an impact crater pushed into the northern coast (carved at 80 % of its radius so a landbridge always survives).
- Water bodies: the classified water promoted into explicit blueprint data (ocean + labeled
lakes with transitional per-body levels —
WBID/WBTB/WSRFsections), plus priority-flood basin diagnostics. TheSkipRoadsconfig toggle turns a full ~26-min generation into a ~80 s road-less iteration cycle. - Logistics:
AStarGrid2Dpathfinding producing a looping continental highway, a branch to the mountain hub, and county roads daisy-chained outward with Prim's algorithm. Paths are decimated (Ramer–Douglas–Peucker) then smoothed (4 Chaikin passes). - Zoning: biome mapping and tiered town/POI placement by slope, temperature and proximity.
⚠️ A* is the known bottleneck
The grid is one node per pixel — 67 million nodes at 8K — and the weight-setup pass walks every
one of them. A full generation can stall long enough that runs get abandoned. await yields are
sprinkled through the road stage, but they cannot interrupt a single engine-side GetPointPath call,
so they only help between paths. Expect a regeneration to take minutes, or to need abandoning.
Outputs — written to user://, not into the project
| File | What it is |
|---|---|
MapData_Seed_<seed>.dat |
The binary blueprint the server reads — v2 tagged container (~336 MB at 8K; contract in Core/Scripts/BLUEPRINT_FORMAT.md) |
MapData_Seed_<seed>_v1.dat |
The same content in the legacy v1 format (~512 MB), dual-written as a safety net until a future removal task |
Map_Seed_<seed>.png |
Visual snapshot of the map, for reviewing and picking seeds |
On Linux: ~/.local/share/godot/app_userdata/islaApocolypse/.
Running it
Open Tools/Scenes/MapPreview.tscn and press F6. It is also the project's main scene, so
pressing F5 anywhere runs this — which is why F5 regenerates your world instead of starting the
game. Use Scenes/MapCaptureTool.tscn when you specifically want the full-resolution PNG.
Rules
- No client dependencies. Nothing here may reference UI, player controllers or shaders in
/Client. - Data out, nothing in. Tools emit
.dat/.png/.json; they never write to a live server save. - No magic numbers. Distances and radii scale off
MapSizeor the derivedscaleFactor, so the generator behaves the same at every profile.