Replace flat-black unexplored fog with an antique-cartography treatment
(Civ-VI style), themed as dwarven 'unmapped vellum/slate' — reveals no
real terrain, only a stylized stone surface.
- fog_renderer.gd: unexplored tiles painted with a procedurally generated
opaque vellum texture (FastNoiseLite domain-warped FBM → warm dark
slate→parchment gradient), generated synchronously at init so there is
no reveal-before-ready leak. No binary art asset required. Visible/seen
paths unchanged; frontier stays a clean opaque edge.
- design-tokens.json -> ui_theme.tres: fog.unexplored 000000ff -> 1a160fff
(warm 'unmapped vellum' tone) so the minimap unexplored cover matches.
Render-verified on apricot (iter_7q proof, fog enabled): undiscovered
renders as warm stone/vellum on map + minimap, no terrain bleed, clean
frontier, lit tiles unaffected. gdlint clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Undiscovered (never-seen) terrain was leaking through fog on both the
main map and minimap; only the lit(visible) vs unlit(seen) distinction
worked. Root cause: unexplored overlay was sub-opaque + undersized.
- fog_renderer.gd: UNEXPLORED_COLOR alpha 0.85 -> 1.0 (opaque); edge-fade
softening now gated to the FOGGED state only — unexplored tiles stay
fully opaque even at vertices bordering a visible tile (was revealing
undiscovered terrain along the exploration frontier).
- design-tokens.json -> regenerated ui_theme.tres: fog.unexplored
alpha 0.90 -> 1.0 (fixed at token source, not hand-edited).
- minimap.gd: unexplored now drawn as a full tile-pitch opaque cover
instead of a 3x3px dot (the dot left ~5px gaps at minimap scale,
leaking terrain). fog.explored (seen dimming) left unchanged.
Verified on apricot via iter_7q_worldmap_visual_proof with fog ENABLED:
undiscovered renders solid black on map + minimap, clean hard frontier,
lit tiles unaffected. (GUT cannot prove this — render-verified.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Full integration of worktree-bridge-cse_01Nnt (54 commits) into main.
Conflicts (17) resolved: adopt branch's p2-65 GameState->mc-state crate move
across mc-ai/mc-core/mc-player-api/mc-turn; keep main's design-token theme
system (build-ui-theme.py, tile_info_panel.gd) and main's newer p0-26b evidence.
Workspace: 2624 tests pass; 5 pre-existing/environmental failures inherited from
main (constructor_smoke capping test, five_players_overflow, 3x gpu_rollout_parity)
— zero regressions introduced by the merge. api-gdext (GDExtension) compiles.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Route inline Color() in the world-map HUD banners and minimap onto ThemeAssets
design tokens. Visual-only (Rail 3), no logic change:
- world_map_hud: patrol banner box -> background.panel/accent.science + label
accent.science; rally picker box -> background.panel/accent.sage + title
accent.science; notification toast box -> background.panel/accent.gold +
label text.primary
- minimap: const FOG_COLOR/UNEXPLORED_COLOR/PING_RING_COLOR -> _ready()-populated
vars from fog.explored/fog.unexplored/accent.ping; border stylebox border ->
border.panel; backdrop fill -> background.deepest
Minimap TERRAIN_COLORS biome palette stays hardcoded with an explicit comment:
it is game-content color (no UI token applies) and a deliberately distinct muted
minimap rendering, not a terrain.json mirror (the arrays diverge). Re-sourcing it
is a Rail-2 data follow-up, out of scope for this visual-only token pass.
DataLoader RGB conversion, env-var viewport color, transparent structural fills,
and the neutral gray default left as-is.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to 268b85c92. The first tokenization mapped the per-state button
styleboxes such that the warm-token luma ordering inverted the interaction: the
button darkened on hover instead of brightening. Reorder by luma so hover is the
lightest warm token (button.bgPressed #472f0f), pressed the darkest
(button.bgHover #331a0d), and normal in between (background.listSelected
#3f2d0d) — restoring the original gold-emphasis brighten-on-hover feel. Token
names read awkwardly because the button.* set targets the indigo-idle inherited
button, not a gold emphasis button; the comment documents the luma intent.
Visual-only; compiles with autoloads (GATE_OK).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Ran the 10-seed T300 autoplay smoke batch on apricot against the committed
build (origin/main @ f5c4ee9c3, the live pick_research_via_bridge dispatch).
Recorded the per-seed tier_peak table as cited evidence: all 5 clans present,
all games resolved to victory with both sides active (seed5's 37-turn win is a
legitimate tinkersmith domination rush), tier_peak spans T1-T6 with no
collapse — a healthy, clan-diverse post-port distribution.
Bullet 5 remains [ ] / status partial: the acceptance is comparative
("unchanged-or-better vs the pre-port baseline") and no recorded pre-port
baseline exists. The nearest pre-port SHA is confounded by ~3 weeks of
unrelated changes; the only confound-free before/after (port commits reverted)
isn't reachable by the apricot launcher (forge-fetch only, revert branch is
ACS-unpushed). Decision-correctness is already pinned by the bullet-4 parity
test (8/8). tier_peak extracted directly from player_stats[].tier_peak because
autoplay-report.py aborted on a pre-existing autoplay-validate.py bug
(unhashable type: list) — left unfixed (out of fence).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Log the 17 tokenized HUD/notification scripts, the three landing commits, the
const-Color-trap conversions, GUT baseline parity (16 pre-existing failures,
zero new), proof paths, and the held-for-world_map remainder (world_map_hud,
weather_visualizer, minimap, encyclopedia). Status stays stub — clusters for
world_map / encyclopedia / minimap remain.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
p2-57b gap analysis flagged dangling produces[] refs: adamantine_press /
gravity_press → dwarf_master_engineer, royal_infirmary / field_hospital_corps
→ master_surgeon, grand_chronicle → saga_writer named units that had no JSON.
Authored all three in public/resources/units/ matching the unit schema and the
shape of their nearest siblings (dwarf_grand_engineer, field_surgeon,
loremaster): Great-Person-tier support specialists with logistics blocks, real
tech gates (deep_engineering / surgical_practice / rune_archives), upgrade
chains, and GP actions consistent with their class. validate-game-data.py
PASSes all three (the 90 unrelated failures pre-exist on main).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
p2-65 (stays partial K=8/9): rewrite Phase-7 resume note. The 'no-persistent-
GdGameState / discards combat_balance' premise is STALE — game_state.gd:269
loads combat_balance.json into the persistent _gd_state singleton via
GdGameState::set_combat_balance_json (api-gdext/src/lib.rs:3541), and the driver
objective p2-55f is already done with its bullet 2 marked [x]. Bullet 9 must NOT
be closed on that path (it bypasses mc-state entirely — would be an
objective-integrity reframe). Remaining work is architectural consolidation
(mc_state::SimConfig owning global config), now an owner decision, not a
mechanical close-out.
p2-72a (stays partial): array-removal increment re-checked and NOT executed.
All three gate-feasibility blockers reproduce — cold Godot import cache, no
mc-godot Docker image (freeze-safe path unbuilt), operator-gated visual proof
unavailable this session. game_state.gd UNTOUCHED (npc_buildings count = 33).
Clean stop re-confirmed; predecessor's execution recipe unchanged.
Regen objectives index. No status frontmatter changed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Read-only analysis of the npc_buildings serialize chain (no code touched;
game_state.gd unchanged). Finding: the resume map's "array removal forces a
SaveManager rewrite, cannot land independent of save-format-migration" claim is
over-stated. `_serialize_npc_buildings` → `serialize_npc_buildings` calls
`b.to_dict()` on each `Building` view, and that view's `to_dict()` already
proxies to `_gd_state.npc_building_dict(_idx)` — serialized bytes are already
sourced from the Rust mirror. Re-pointing serialize at the mirror directly is
byte-identical; the save FORMAT does not change. This objective owns the whole
increment.
STOPPED this session deliberately (gate-feasibility, not coupling): the increment
edits the highest-regression live-world-map file and must land as one green commit
gated by GUT-headless + cargo + build:gdext + operator-gated world-map visual
proof. Godot import cache is cold (host-freeze risk per godot-docker.sh), no
mc-godot Docker image / warm volume exists, and the visual proof needs operator
sign-off — not completable safely without a runway session. Precise execution
recipe (serialize re-point + 6 reader retargets + array/spatial-index/view-fn
deletion + gate steps) recorded for the next session. Status stays partial.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
