beezm
03fe75b378
The bare TerrainGenConfig defaults did NOT reproduce the terrain the developer
locked, so "run the default generator" was not "the terrain in the gallery" —
the single most expensive fact in the codebase, and the reason a fresh chat
would spend an afternoon chasing differences that were CONFIGURATION, not
regression. This is the deliberate task that ends that, before any river work.
A1 — the defaults ARE the locked shape now. Five fields actually move:
SpeckRevert false->true, MinLandComponentFrac 3e-5->2.5e-7 (120x smaller; the
config default would have eaten real islands, not specks), SouthStretch 0->2,
FragmentAmp 0->0.5, Erosion false->true. Seven more were already correct via
SouthernStretch.Default* / CoastalFragment.Default* and are now pinned as
literals, because TerrainShapeV1 used to do that pinning and this default set
inherits the job. CoastShelf stays OFF — the locked shape has no shelf, and
evaluating it (D-041) is its own later task once water renders. Offshore stays
Off permanently (D-063): islands are organic-only, made by the stretch +
fragmentation and identified by the region layer, never placed.
A2 — the preserve mechanism. The curve knots are percentiles of the FAMILY-OFF
land distribution; flipping the defaults would have moved the pool, the knots,
and with them the render field of every batch including terrain-shape-v1
itself. So the pool is pinned family-off (TerrainGenConfig.WithFamilyOff /
CalibrationPool) rather than the knots being baked: calibration stays live, its
INPUT distribution is held still. The pin was a no-op by construction — it sets
the values the defaults carried the instant before the flip — and re-measuring
after confirms it: pool, all six knots, per-seed spread, shaped max,
monotonicity spikeMax and all seven band shares identical.
Applied wider than "in CalibrateCurve": OffshoreIslandsTool,
RegionLabelingTool and SouthernStretchTool generate their own family-off
field for the Phase-1 anchor, so the pool pin alone would NOT have covered
them and their a1 would have failed for a configuration reason. TerrainGenTool
too — it AUTHORED 02_pass1_port and must stay able to regenerate its own
anchor.
Recorded as a judged-and-parked property: knots measured family-off, applied
family-on. Deliberate, not an oversight. Same disposition as the mid-slope
feather.
A4 — no oracle may pass against a superseded baseline. Six anchors retired
(01/03/04/06/08/09) with their checks and ISLA_T0x_SOURCE defaults; three kept
(chat1/02_pass1_port as the family-off pass-1 guard, chat2/10 and chat2/11 as
the shape and erosion acceptance anchors). Two invariants were RE-POINTED
rather than lost — the southern stretch's north-lock and the coastal-fragment
interior-lock now compare against SAME-RUN fields, which is scale-free and
cannot be invalidated by a moved dump. The retired dumps are kept, not deleted,
and marked superseded in their INDEX.md.
A missing anchor is now LOUD. The old pattern skipped silently, so a moved
anchor did not make its oracle fail — it made it not RUN, and a batch with a
skipped check prints an all-PASS table that reads like a clean one. That is
the INVERSE of the hazard the re-baseline guards against, and the migration
below is exactly the event that would have triggered it, on nine anchors at
once. ShapingOracle.LoadAnchor now separates the two cases: absent -> throw;
present at another size -> loud INCONCLUSIVE, which is a fail, never a pass.
TerrainShapeV1 inverted from PRESET to GUARD and moved to its own file.
Apply() is gone — stamping the values on top of the defaults would MASK a
drift instead of catching it. Its constants are now the assertion target, and
Assert() refuses a run whose defaults have drifted off the locked shape.
B/C — batches are namespaced by chat: batches/<chat>/NN_slug/. Task numbers
restart at 00 per chat, so a flat root collided the moment a second chat
existed — four colliding prefixes across 25 batches, separable only by slug.
ToolingPaths.ChatSlug is REQUIRED (throws if unset) and defaults per tool to
its authoring chat, so re-running reproduces a batch in place while ISLA_CHAT
redirects — which is also what stops an acceptance run from overwriting the
very anchor it checks against. Writes go through BatchRoot; historical READS
compose against BatchesRoot and so carry the prefix in their own source string
("chat1/02_pass1_port"). The 25 existing batches were migrated moves-only.
ACCEPTANCE — 16 of 16 byte-identical, 0 failed. All 8 gallery seeds at 8192
from the bare defaults are byte-identical to chat2/10_frag4_seed_gallery
(= terrain-shape-v1,
|
||
|---|---|---|
| Client | ||
| Core | ||
| Server | ||
| Tools | ||
| .gitattributes | ||
| .gitignore | ||
| islaApocalypse-v2.csproj | ||
| islaApocalypse-v2.sln | ||
| project.godot | ||
| README.md | ||
Isla Apocalypse — v2
The rewrite (D-049). A post-apocalyptic survival voxel game: one island, generated once offline, loaded by everyone.
Status: Phase 2 — shaping. Phase 0 laid the contracts and the tooling rails; Phase 1 ported the crown-jewel noise and pass 1 (the island's raw shape); chat2/01 added pass 2a — the height redistribution curve and shelf detail, re-calibrated against this repo's own output, with the two-height-field split and an automatic oracle.
Still no erosion, rivers, water bodies, crater carve, coast shelf, offshore islets, biomes, roads or mesher.
⚠ Two repos. Never one.
| Repo | Role |
|---|---|
~/celerNexus/islaApocalypse-v2/ (this) |
The rewrite. Every line of new code is written here. |
~/celerNexus/islaApocalypse/ |
⛔ READ-ONLY. The pre-rewrite prototype, archived at ab78883, tag pre-rewrite-reference. Ported FROM; never written to. |
~/celerNexus/islaApocalypse-vault-v0.1/ |
The design vault — the source of truth for WHAT and WHY. The vault leads. |
The old codebase is a spent quarry (D-050): mined, never tracked, never synced, never canon'd. Writing rewrite code into it would corrupt the very thing being ported from.
⚠ Runtime data isolation — do not weaken
This project pins its user:// directory away from the old prototype's, which holds the
preserved reference seeds, blueprints and batches:
user:// → ~/.local/share/islaApocalypse-v2/ ← this project
~/.local/share/godot/app_userdata/islaApocalypse/ ← ⛔ the OLD prototype. Never touch.
Pinned two ways in project.godot — a distinct application/config/name and
use_custom_user_dir + custom_user_dir_name. Both must stay.
⚠ Do not document this with a comment in
project.godot— it will not survive.Godot rewrites
project.godoton import and STRIPS ALL COMMENTS, replacing the header with its own boilerplate. (Measured: a warning block written there was gone after the next--import. The settings survived; only the explanation was deleted.) So the warning lives here, and the guarantee is enforced by the probe below — which exits non-zero on a clash and therefore works in a pre-flight check, not just by eye.
After any change to the [application] block, re-run the probe:
Godot_v4.7.2-stable_mono_linux.x86_64 --headless \
--path ~/celerNexus/islaApocalypse-v2 res://Tools/Scenes/UserDirProbe.tscn
Layout
Core/ math + data only, ENGINE-FREE — depends on nothing above it
Server/ authoritative logic — may use Core (empty: not yet its phase)
Client/ rendering — may use Core (empty: not yet its phase)
Tools/ the offline generator — ⚠ may NOT use Client
Each layer has a README.md stating its role and its boundary. Read the one for the layer you are
about to write in — they carry the reasons, not just the rules.
The contracts this phase establishes
- ★★ The column model (D-053) — the world as a 2D grid of columns, each a stack of
(material, thickness)runs. Air is the top run; there is no air-vs-solid height branch. →Core/Scripts/Column.cs - The material schemas (D-054) — terrain and building, two schemas bridged by a transformation
seam that is deliberately empty. Identity is a registry key, never an ordinal. Mesh style is
per-origin, never a material property. →
Core/Scripts/MaterialRegistry.cs - The scaling discipline — nothing uses a raw pixel number.
→
Core/Scripts/GenerationScale.cs - The vertical yardstick — ONE metres-per-raw-unit number, not three constants and twenty
literals as the prototype had. →
Core/Scripts/WorldScale.cs - The elevation profile — the 7-band redistribution curve, identity at and below sea, its
knots measured rather than chosen. →
Core/Scripts/HeightCurve.cs - The two-height-field split (D-046) —
heightClassifyis raw and is the oracle;heightis curved and is what gets drawn. →Tools/Scripts/Pass2Result.cs - Tooling safety — every path env-overridable, deletion refused by code.
→
Core/Scripts/ToolingPaths.cs,Core/Scripts/FileSafety.cs
Build
dotnet build
Godot_v4.7.2-stable_mono_linux.x86_64 --headless --import --path .
Godot 4.7.2 (mono) · .NET 8 (net8.0) · Godot.NET.Sdk/4.7.2
The reference repo is Godot 4.7.1. This repo deliberately targets 4.7.2 — clean rewrite, current tooling. 4.7.1 is the port source, not a constraint on new code.
⚠
<RollForward>Major</RollForward>in the.csprojis load-bearing.The target is
net8.0but this machine carries only a .NET 10 runtime. Without that property the generatedruntimeconfig.jsonrolls forward by MINOR only and every run fails with "You must install .NET 8". This bites at run time, not build time — the build succeeds either way — so re-check the emittedruntimeconfig.jsonafter any.csprojchange.
What this phase is NOT
No mesher, no stratigraphy, no biomes, no water, no rivers, no erosion. The pipeline is 2D-maps-first (D-056): stages 1–5 are flat maps, gotten completely right before a single 3D vertex exists. Building the mesher before the data is right is the specific trap this rewrite exists to undo.
And the curve is a BASELINE, not a verdict. It is ported faithfully and calibrated honestly so that later reshaping has a control to be judged against. Whether this elevation profile — 60 % of land in the bottom 14 m, 4 % above the plateau — is the one the world wants is an open question the histograms exist to inform, not one this port has answered.