feedback5: merge feedback4 (3af5b84) — PRs #186 / #134 / #185 / #24 / #183 ported to the split layout (terrainInfo / TERRAIN_TIPS into a new ui/gameLogic/terrain.js re-exported by the gameLogic.js barrel, closesOnFieldPress into ui/gameLogic/shortcuts.js, the rest in place; 42 strings through t(), #183's 设置 too); golden unchanged

Ported by hand:
- public/js/ui/gameLogic.js is a re-export shim on feedback5: #185's TERRAIN_TIPS / terrainInfo went into the new
  public/js/ui/gameLogic/terrain.js (the barrel gains one re-export line); closesOnFieldPress 'terrain' into
  ui/gameLogic/shortcuts.js.
- i18n (docs/I18N.md): every new user-visible string goes through t() with an English entry in public/i18n/en.json
  (42, after the 补位 strings): the terrain names / tags / fact table are N_() and translated where terrainInfo hands
  them out, the mechanism lines are msgids with their numbers as params (a line with an optional half is two whole
  msgids, the list joins use an array param), the card's 地形机制 / 这一格 titles in ui/detailPanel.js. Official EN
  terms where the EN client has them (stage_table tileInfo: Active Originium, Mire, Exhaust Grille, Deep Water Zone,
  Incursion Point, Protection Objective, Tunnel Entry / Exit, High Ground; Camouflage). The Chinese output is
  byte-identical to feedback4's terrainInfo on all 5187 tiles of the 13 stages and on synthetic stages.
- public/js/screens/game.js (MatchScreen still lives there): the import keeps feedback5's names plus terrainInfo /
  fieldTile; the terrain stage is the one view.setStage gets, so it includes feedback5's fieldStage — a 联防 field is
  played on its own map (GitHub #41), and a tap there explains that map's tile, not the round stage's. The tileClick
  handler and render/app.js emitTileClick / ui/fallbackField.js auto-merged in place. terrainInfo reads the stage only,
  so a 补位 / 自选 board explains its tiles the same way (their pieces only change a piece's art and name).
- public/js/ui/detailPanel.js: the terrain branch of resolveDetail before feedback5's stand-in / 自选 helpers.
- public/js/screens/title.js (#183): feedback5's LangToggle / t() imports plus GIcon / SettingsModal; the gear
  button's aria-label / title are t('设置') (en.json already has "Settings").
- tools/fetch-assets.mjs / tools/assets/downloader.mjs: --add-only (keepExisting, the font skip) beside #24's
  --asset-source / SP_ASSET_SOURCE and the shared MirrorPolicy; feedback5's extraOperators / moduleTypes and the
  voice plan unchanged.
- data/assets.json: feedback5's manifest plus #134's audio.sfx.battle.leak (after heal, the plan order); stats files
  6993 → 6994 and bytes +23 240 as measured on feedback4; the hash recomputed with the generator's contentHash
  (c01aa739a2d3 → 6a547fc5c937). A field-wise JSON diff against feedback5 shows only those four fields.
- docs: DESIGN §24 intro and §24.5 / §24.6 as on feedback4 (the rest of DESIGN.md is feedback5's); ASSETS.md,
  DEPLOY.md, PLAYING.md auto-merged.
- Auto-merged and checked: audio.js (the sim still emits ['leak', id], server/sim/battle/deploy.js), render/style.js,
  textures.js, tiles.js, board3d/materials.js, boardArt.js, tools/assets/*, tools/setup.mjs, extract.py, the tests.

Checks: the PRs' unit tests (audio 22, gameLogic 44, terrain-detail 1, originium-art 4, boardart 15,
board3d-extract 9, asset-network 21, fetch-assets-shrink 4), i18n 9 / i18n-data 8 / i18n-server 4, client-static 313,
docs-consistency 24, kits_layout 6 pass; lint 0 errors (no new warnings), check:imports, typecheck; golden: all 283
scenarios match, nothing regenerated.

docs/DESIGN.md

@@ -2172,7 +2172,7 @@ Decisions of the owner, 2026-10-04: a 联防 devour counts members who enter alr
 
 ## 24. Community reports after 0.1.3
 
-Reports after the 0.1.3 release. Each was checked against the official data and PRTS; anything with no source is marked [ASSUMED]. Where each is handled: a skill clip with no Begin and no own Idle plays once (德克萨斯 S2 剑雨) → §24.1; 高台 for every melee chess whose trait reads 「可以放置于远程位」 (GitHub #153, PR #69) → §24.2; the 5-阿戈尔 revives go to the first 3 members knocked out (GitHub #105, #140) → §24.3; 联防 阿戈尔 devours the operator in front whoever owns it (GitHub #140) → §24.4. Merged pull requests keep their own notes where they changed the rules: deploy-timed skills as duration skills (PR #109, §5 skill contract), the teammate's hand / temp / equipment in prep scouting (PR #129), 灵巧's knocked-out aura and the hidden-layer bonds (PR #66, research 02).
+Reports after the 0.1.3 release. Each was checked against the official data and PRTS; anything with no source is marked [ASSUMED]. Where each is handled: a skill clip with no Begin and no own Idle plays once (德克萨斯 S2 剑雨) → §24.1; 高台 for every melee chess whose trait reads 「可以放置于远程位」 (GitHub #153, PR #69) → §24.2; the 5-阿戈尔 revives go to the first 3 members knocked out (GitHub #105, #140) → §24.3; 联防 阿戈尔 devours the operator in front whoever owns it (GitHub #140) → §24.4; a special terrain tile explains itself on a tap (GitHub #184, PR #185) → §24.5; 活性源石 was two different materials on the two boards (GitHub #184, PR #185) → §24.6. Merged pull requests keep their own notes where they changed the rules: deploy-timed skills as duration skills (PR #109, §5 skill contract), the teammate's hand / temp / equipment in prep scouting (PR #129), 灵巧's knocked-out aura and the hidden-layer bonds (PR #66, research 02).
 
 ### 24.1 A skill clip with no Begin and no own Idle plays once — 德克萨斯 S2 剑雨 (player report) — `render/spine.js SpineActor.setSkill / update` (`SKILL_CLIP_MIN`, `skillClipOnce`, `skillEndPending`)
 
@@ -2342,3 +2342,23 @@ Reports after the 0.1.3 release. Each was checked against the official data and
   through a teammate's 幽灵鲨 equals the own chain, and her knock-out takes her owner's 5-tier slot; the own-board case
   through `uniteBattleOpts` (carry.down from a real own combat, the own combat's gains). The first three fail on the
   previous code. `feedback3-egir-down`, `bonds_core`, `unitedown`, `feedback3-unitecarry` pass unchanged.
+
+### 24.5 A special terrain tile explains itself on a tap (GitHub #184, player request 「建议加入对于特殊地形的单击信息提示」) — `render/app.js` (`tileClick`, `emitTileClick`), `ui/fallbackField.js` (the board tile), `ui/screens/game.js` (`tileClick` → `terrainInfo`, `live.terrainStage` / `terrainTile`), `ui/gameLogic.js` (`terrainInfo`, `TERRAIN_TIPS`), `ui/detailPanel.js` (`TerrainDetail`, `resolveDetail`)
+
+- **Asked** (GitHub #184, after 0.1.3): a tap on a special terrain tile should say what it is. The mode draws 活性源石, 沼泽, 排气格栅, 深水区, the gates and the teleports, but nothing in the client names them — the legend lived in `docs/DATA.md` and the data only.
+- **Why the ground was free**: a press on a tile that had nothing on it only closed what was open (a piece / unit card, the selection). `render/app.js` now emits `tileClick { row, col }` when neither a battle unit, a prep piece (the drag controller), a pen enemy nor the boss leader took the press, and `ui/fallbackField.js` emits the same from its own tile divs, so both field implementations behave alike.
+- **What the card says**: `gameLogic.terrainInfo(stage, row, col)` reads the glyph from `stage.rows`, its legend entry from `stage.tiles` (tileKey / height / buildable / passability) and **that stage's own terrain parameters** from `stage.special[<terrain>]` — the values `sim/content/devices.js` runs on: 活性源石 `damage 70 / atk 0.2 / attack_speed 20 / duration 300` (PRTS 特殊地形 "部署于其上的我军和经过的敌军在{duration}s内每秒受到{damage}真实伤害…"), 沼泽's stack rule (`intervalSec` 1, `aspdPerStack` −0.05, `moveMulPerStack` −0.05, `maxStacks` 10, `heavyWeight` 3, cleared on leaving — PRTS 沼泽控制), 深水区's drowning (`damage 40`, aspd −0.6, move ×0.6) plus 拒绝部署 (PRTS 深水区 地形信息; `grid.js DEPLOY_REFUSED_TILES`), 排气格栅's "operators on it cannot be targeted by enemy ranged attacks" — completed in review with the two halves that matter in play: the operator it blocks still gets attacked (the tile's buff is a 隐匿-like `stealth` flag, `devices.js enterTerrain`), and 深水区's drowning damage is sourceless damage that no 干员's 增伤 / 穿透 / 装备 applies to and is *not* 环境伤害 (`tickDeepsea`'s tags are `dot` / `periodic` / `deepsea`, deliberately not `terrain`). The gates and teleports carry no `special`, so they key on `tileKey` — 红门 (enemy gate) / 蓝门 (protection objective) / 传送入口 / 传送出口, in the words `docs/PLAYING.md` already uses. The card is a normal `ResultDialog`-less `DetailPanel` card: read-only, no protocol frame, closed by the next field press (`closesOnFieldPress` now lists `terrain`) or its own ✕.
+- **An ordinary tile still says nothing** (road / floor / wall / fence / bench → null): the press keeps its other meanings there — deselecting a piece, closing a card — and a drag still starts only on a piece.
+- **Coordinate spaces (corrected in review)**: a tap reports BOARD coordinates — `emitTileClick` picks with `pickBoardTile`, so the boss-prep transform (`prepXf.toBoard`) is applied once, in the renderer. (The first version reported the DRAWN tile from `groundTile` and the screen mapped it again, so on a Final Assault / Hidden Core prep the tap was converted twice and explained another tile.) Everywhere but that prep board the two spaces are the same: a 最终攻势 / 隐秘核心 battle renders the stage's own rows (`GEO.BOSS_RECT`), the normal and 联防 rects are stage rows. A boss PREP draws the player's half — stage rows 2–5 — as board rows 9–12, and the screen turns the board tile into the stage tile with `gameLogic.fieldTile` before the lookup (and reads the stage the board on screen is built from: the own one with the player's 机变 tile overrides, `effectiveStage`, or the plain stage while watching a teammate).
+- **[ASSUMED]**: that a player wants the official's 地形信息 here (the official shows terrain descriptions in its own level / terrain UI, not in-match — research 09 records none) and the wording of the mechanism lines: the numbers are the data's, the prose is ours (marked as ours in `TERRAIN_TIPS`). A 机变 card that switches a terrain off (模拟战场演变 · 沙尘暴停止) is not reflected — the tip explains the TILE, not the round's modifier.
+- **Tests**: `test/ui/gameLogic.test.js` (every terrain of the real stages with its own numbers: 活性源石 in 战场#04, 沼泽 in 战场#06, 排气格栅 in 战场#07, 深水区 in 战场#05; the four gate / teleport tiles; an ordinary glyph → null; out-of-range, fractional and legend-less input; the fact lines from the legend entry; a stage missing its parameters still explains the mechanism), `test/ui/terrain-detail.test.js` (the tap's target resolves into the card; a tip-less or malformed one does not), `test/ui/terrain-tip.e2e.test.js` (headless Chrome on the mock harness: a tap on 蓝门 / 红门 opens the card with its own line on BOTH field implementations — the engine canvas at the tile's own screen position and the DOM fallback's tile — while a tap on a floor tile opens nothing; and on a Final Assault PREP the tapped tile is the one the board tile maps to on the boss field, the regression the review asked for).
+
+### 24.6 活性源石 is ONE material on both boards (GitHub #184, same report's second half) — `render/style.js` (`ORIGINIUM`, `linearToSrgb255` / `linearToHex`), `render/textures.js` (`originiumCanvas` / `originiumOverlay`, `MAT_DRAW.infection` / `infection2`, `PROC_LAYERS.originium` / `originium2`, `DEFAULT_ART_LAYERS.infection` / `infection2`), `render/tiles.js` (`variantMat`, the pulse tint), `render/board3d/materials.js` (`infectionMaterial`)
+
+- **Reported** (GitHub #184): 「感觉源石地板的图片有点奇怪（有点简陋，另外两个源石地板的贴图不一致）」 — two originium floors that do not match. They were two different materials: with the board art installed (or `?board=2d`) the tile came from the 2D atlas cell — a beveled brick with seven randomly placed orange crystal clusters — while the 3D board drew a world-space crust with pulsing veins over the official concrete. And that 3D crust was multiplied by a per-tile fade (`edge = smoothstep(0, 0.1, min(vUv, 1 − vUv))`), so every tile of a 活性源石 area faded out at its own border and the area read as separate patches rather than one floor.
+- **One palette**: `style.js ORIGINIUM` holds the crust's dark and lit sides, the vein and the crystal grains. The values are the 3D shader's own working colours (its fragment ends in `#include <colorspace_fragment>`, so they are linear); the canvas converts them with `linearToSrgb255` — without that the 2D tile came out darker than the 3D one, which is what the first version of this change did.
+- **One recipe**: two octaves of noise make the crust mask (`smoothstep(0.38, 0.62, n·0.7 + n2·0.45)`), the ridge of the fine octave its veins (`smoothstep(0.86, 0.98, 1 − |2·n2 − 1|)`, brightened where the crust is), the rare top of a third octave the grains — the shader runs it over world-space noise, `textures.js originiumCanvas` over a *toroidal* field (the noise wraps, so a cell that is repeated over the board lines up with itself: the veins run through the tile's edge instead of stopping at it).
+- **Both reach the tile's edges**: the 3D material no longer has the per-tile fade, and the 2D cell has no frame any more (the old `bevelTile` is gone). The mire / water / smog materials keep their own fades — this section is about 活性源石, the tile that was reported.
+- **With the board art installed** the 2D tile is composed exactly like the 3D one: `DEFAULT_ART_LAYERS.infection` draws the official concrete slab (`board3d/atlas.js concrete`, the same atlas rect) and then the crust as a procedural layer; without the art the cell falls back to `MAT_DRAW.infection` (the procedural ground + the same crust). A second variant (`infection2`, the same art mirrored; `PROC_LAYERS.originium2` for the art path) is picked by tile (`tiles.js variantMat`), so a field of 活性源石 does not repeat one pattern.
+- **Not [ASSUMED]** — this is our own art: the palette and the recipe are ours, the tile it dresses and its numbers are the data's (a 机变 card can switch the terrain off, §24.5).
+- **Tests**: `test/render/originium-art.test.js` (the couplings: both renderers take the palette from `ORIGINIUM`, the canvas converts it, the recipe's thresholds agree, the 2D base slab is the same atlas rect the 3D board lays its crust on, the infection shader has no per-tile fade while the mire keeps its own, two variants are registered and picked, the pulse tint is the shared one), `test/render/originium-cell.e2e.test.js` (headless Chrome: the cell's veins hit the palette's displayed colour (`linearToHex(ORIGINIUM.vein)`), veins reach the border like the middle (no frame), the veins at opposite edges are alike (the cell joins itself), and the second variant is a different pattern).