Resolves the divergence-register conflict: kept the accurate per-VERTEX AP-35
(Fix A shipped per-vertex; main's row was the stale pre-Fix-A per-pixel text),
kept main's UI rows AP-37..AP-42, and renumbered this branch's torch-gate row
AP-37 -> AP-43 (AP-37 was taken by main's LayoutDesc row). AP count 41 -> 42.
Retargeted the AP-37 references in WbDrawDispatcher + the CHECKPOINT to AP-43.
Marked ISSUES #140 RESOLVED (b7d655b) with the corrected root cause.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Holtburg meeting-hall facade washed out warm/bright vs retail. The round-1
checkpoint blamed torch REACH (acdream Falloff 6×1.3=7.8m vs a supposed retail
Falloff 4). That theory is WRONG, and this commit fixes the real cause.
Empirical (HoltburgTorchFalloffProbeTests, headless dat dump via the production
LightInfoLoader): the orange entrance torch (setup 0x020005D8) is raw dat
Falloff 6 and acdream reads it FAITHFULLY — there is no Falloff-4 torch anywhere
in Holtburg. Both clients read the same dat float, so reach was never inflated.
Decomp (read verbatim + corroborated by an independent adversarial workflow):
retail's per-object torch binder minimize_object_lighting (0x0054d480) is gated
in RenderDeviceD3D::DrawMeshInternal (0x0059f398) by `if (Render::useSunlight == 0)`.
The outdoor landscape stage runs useSunlightSet(1) (PView::DrawCells 0x005a485a,
before LScape::draw), so the building EXTERIOR shell — drawn via
DrawBlock→DrawSortCell→DrawBuilding→CPhysicsPart::Draw→DrawMeshInternal — is lit
by SUN + ambient ONLY; torches are SKIPPED. The static bake
(SetStaticLightingVertexColors 0x0059cfe0) is EnvCell-only. So retail NEVER
torch-lights outdoor objects. This exactly explains the isolation test (object
point lights OFF → building matches retail).
Fix: WbDrawDispatcher.ComputeEntityLightSet gates per-object torch selection on
the object being INDOOR (ParentCellId is an EnvCell, (id&0xFFFF)>=0x0100) via the
pure predicate IndoorObjectReceivesTorches. Outdoor objects (building shells with
null ParentCellId, outdoor scenery, outdoor creatures) keep the all-(-1) light
set ⇒ sun + ambient only = retail. The indoor "no sun" half is already handled by
the global sun-kill when the player is inside a cell (UpdateSunFromSky). No
dungeon regression: EnvCell statics get ParentCellId set (keep torches).
Divergence register: AP-37 (residual: acdream keys sun/torch on the object's own
cell + a per-frame player-inside sun-kill, vs retail's per-draw-stage useSunlight;
only matters for through-doorway look-ins). The round-1 CHECKPOINT got a RESOLVED
banner correcting the reach theory.
Tests: WbDrawDispatcherTorchGateTests (7), HoltburgTorchFalloffProbeTests (dat
dump). App 280/1skip, Core 1486/2skip green. Held at the visual gate — not merged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Same-instant cdb proved acdream ambient (0.447) == retail (0.4465) and time/sun match,
so the building/character over-brightness is NOT the bake/wrap/EnvCell/clamp (D-1..D-4,
all correct but off-target) — those light the wrong surfaces. The Holtburg building
exterior is a mode-0 OBJECT (IsBuildingShell, not an EnvCell). Isolation (object point
lights gated OFF) made it match retail => cause is the torch REACH being too long
(acdream range 7.8 = Falloff 6x1.3 vs retail 5.2 = Falloff 4x1.3), flooding the small
facade. OPEN: confirm same-torch Falloff acdream-vs-retail before tightening the reach.
Diagnostic shader hack reverted (tree clean); D-1..D-4 kept. Branch not merged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mesh_modern unified all meshes into one calc_point_light path: it applied the
bake's half-Lambert wrap to objects (lighting character backs from a torch behind
them) and added the sun to EnvCell building shells (warm facade wash). Retail
splits these: objects = hardware plain Lambert max(0,N.L) + sun; EnvCell walls =
baked wrap, dynamics only, NO sun (minimize_envcell_lighting). Add a per-draw
uLightingMode (WbDrawDispatcher=0 object, EnvCellRenderer=1 envcell) selecting the
angular term (wrap vs plain Lambert) and gating the sun. Per-light cap + D-1 clamp
unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
AP-35: the "numerically equivalent" claim was false. Residual is now two
parts: (a) per-frame GPU evaluate vs retail's bake-once (architecture/perf
difference only; formula matches), and (b) SelectForObject 8-cap means a
surface reached by >8 point lights is dimmer than retail's uncapped bake.
Cross-references AP-16 for the cap ownership.
AP-16: the old "global nearest-8 viewer-distance into UBO" description was
stale — the UBO point-light path is now vestigial (mesh_modern.vert skips
posAndKind.w!=0 entries; point lights come exclusively from the per-object
SSBO binding 5). Retargeted to the current SelectForObject per-object/cell
8-cap mechanism with correct file:line (LightManager.cs:234), both call
sites (ComputeEntityLightSet + GetCellLightSet), and the retail oracle
distinction (hardware cap 0x0054d480 faithful; bake 0x0059cfe0 not).
Preserved the UBO-directional-only note inline rather than losing it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fix A (aa94ced) moved point lighting to per-vertex Gouraud and ported the
half-Lambert wrap + norm distance attenuation. Fix D D-1 added the separate
point-light accumulator clamped to [0,1] matching retail's
SetStaticLightingVertexColors bake clamp.
AP-35 previously stated the path was per-pixel (mesh_modern.frag:52) and
that wrap + normalization factor were "neither ported" — both wrong. Rewrite
to reflect current state: per-vertex in mesh_modern.vert (pointContribution),
wrap + norm ported, point sum clamped. Residual is architecture-only (per-
frame GPU evaluate vs retail bake-once), not a visual divergence.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The cell shell read whatever light set (SSBO 4/5) WbDrawDispatcher last left
bound, lighting walls with a leaked set. EnvCellRenderer now uploads its own
binding=4 global lights (frame PointSnapshot via GlobalLightPacker) + a binding=5
per-instance set, computed per cell by LightManager.SelectForObject over the
cell's world bounds (mirrors _cellIdToSlot + WbDrawDispatcher.ComputeEntityLightSet).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
accumulateLights folded ambient+sun+torches into one accumulator clamped only
in the frag, so a few warm intensity-100 torches blew walls/objects to white.
Mirror retail SetStaticLightingVertexColors: sum point/spot into pointAcc, clamp
to [0,1] (the baked emissive), THEN add ambient+sun, frag final-clamps. Matches
LightBake.ComputeVertexColor (LightBakeConformanceTests). Per-light cap unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Task-by-task TDD plan: (1) extract GlobalLightPacker (Core, pure) + test + refactor
WbDrawDispatcher; (2) lock the bake contract via LightBake conformance test on the
captured golden torches; (3) D-1 clamp the point-light sum on its own in
mesh_modern.vert; (4) D-2 EnvCellRenderer binds its own per-cell light set (SSBO 4+5)
via SelectForObject over cell bounds; (5) correct register AP-35 + reconcile Fix B.
Concrete code + exact insertion points; visual verification is the acceptance gate.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Resolve the Fix D contradiction with decomp (workflow wf_f660eb88 + adversarial
verify) + 4 live cdb captures. The D3D-FF model was the WRONG oracle: retail has
TWO light systems — STATIC torches BAKE into wall vertices (calc_point_light,
triple-clamped: range gate + per-channel min(scale*color,color) + per-vertex
[0,1] from black), DYNAMIC lights go D3D hardware. The captured intensity=100 is
the purple PORTAL (magenta, dynamic), not a wall torch. Ground truth: 38 static
warm torches (orange (1,0.588,0.314)/cream, intensity=100, falloff 3-5) + 2 dynamic.
acdream over-brightness = two confirmed bugs: D-1 mesh_modern.vert folds
ambient+sun+torches into one UNCLAMPED accumulator (single frag clamp) -> warm
blowout; D-2 EnvCellRenderer never binds SSBO 4/5 so the cell shell reads a leaked
light set. Spec: D-1 in-shader clamp-split (clamp the torch sum on its own before
ambient/sun); D-2 bind the shell's own per-cell light set (mirror WbDrawDispatcher);
LightBake.cs is the C# conformance oracle. Adds the 4 reusable cdb capture scripts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A7 lighting Fix A/B/C shipped this session; Fix D (object torch over-brightness)
grounded but blocked on the render-path capture. Filed as #140 + divergence
register UN-7 (object point-light model unconfirmed). Detail in the 2026-06-18
handoff doc.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Brings Fix C (57c1135, sun-vector magnitude / ~32% over-bright) + the A7 lighting
handoff doc onto main. Auto-merged clean against the D.2b line. Merged tree builds
green; 18/18 sky tests pass. Fix A/B already on main (37911ed).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Session handoff: live-cdb grounding shipped Fix A (point-light shape), Fix B
(per-object selection), Fix C (sun-vector magnitude / ~32% over-bright). Fix D
(outdoor objects too bright near torches) is fully grounded but BLOCKED on one
capture (the building's render path) — the D3D-FF math says it'd make objects
brighter, so not ported. Full cdb cheat-sheet + the contradiction + the next
capture in the doc.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Outdoor lighting was ~32% too bright (washed-out, weak shading). Live cdb on
retail (SmartBox::SetWorldAmbientLight + SkyDesc::GetLighting + LScape::sunlight,
binary matches refs/acclient.pdb) pinned it: at the SAME game time + DayGroup,
acdream's ambient COLOR matched retail exactly (the purple is correct, authored
per-time-of-day in the sky dat) but the LEVEL was 0.607 vs retail's 0.459.
level = AmbBright + 0.2·|sunVec|, both AmbBright=0.40, so acdream's |sunVec|≈1.06
vs retail's ≈0.30. Retail's LScape::sunlight read live = (0.2238, ~0, 0.00352),
magnitude 0.224 = DirBright, y≈0.
RetailSunVector had `y = cos(P)` (≈1) — the raw PRE-transform value SkyDesc::
GetLighting writes to arg5 (0x00500ac9), before LScape::set_sky_position's
world transform. acdream ported the un-transformed vector, so the y=cos(P)≈1
term inflated |sunVec| to ~1.06. That magnitude feeds BOTH the ambient boost
(SkyKeyframe.AmbientColor) AND the sun colour (SkyKeyframe.SunColor =
DirColor×|sunVec|), over-brightening the whole scene (terrain, objects, sky)
~30% and also pointing the sun the wrong way.
Fix: RetailSunVector = DirBright × (cos(P)·sin(H), cos(P)·cos(H), sin(P)) — the
world-space spherical form LScape::sunlight actually holds; |sunVec| == DirBright
for all H/P. After: acdream ambient (0.353,0.176,0.449) vs retail (0.360,0.180,
0.459) — within ~2%, user-confirmed "better outside". Sun direction also corrected
(was pointing ~North from the bad y term).
Tests updated to the cdb-verified values (the prior tests pinned the inflated
magnitude). 18/18 sky tests green. reference-retail-ambient-values memory updated.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Deferred cosmetic polish after the widget-generalization landing: tune the
per-ChatKind transcript text colors against retail, and add pressed/hover state
feedback to the chat buttons (UiButton draws only its default state today; the
dat carries Normal/Pressed/Highlight). Not a regression — the generalized chat
matches the prior hand-made build (user-confirmed).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Record the D.2b widget-generalization landing: generic Type-registered widgets
built by DatWidgetFactory, thin find-by-id controllers, the ConsumesDatChildren
leaf rule, Type-3-not-registered decision, and the centered-UiText vitals numbers.
Both visual gates user-confirmed; 404 tests green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The vitals cur/max numbers now render through the generic UiText widget — retail
gmVitalsUI uses UIElement_Text for them, not a meter-internal label. VitalsController
attaches a centered, non-interactive UiText child to each meter and stops the meter
drawing its own label (UiMeter.Label -> null). New UiText.Centered draws the first line
centered H+V with the SAME formula UiMeter's overlay used, so the numbers are
pixel-identical — user-confirmed in the live client.
This completes the D.2b widget-generalization pass: every chat + vitals widget is now
built generically and registered to its retail Type (Button/Field*/Menu/Meter/Scrollbar/
Text), with thin find-by-id controllers. (*Field is controller-placed; Type 3 stays
UiDatElement for chrome.)
Divergence register: AP-37 vitals-numbers-via-UiMeter.Label clause retired. Full suite:
404 passed, 2 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The generalized channel menu wouldn't open: the factory recursed the Type-6
menu element's dat children, building its invisible Type-12 label child as a
UiText. Hit-testing is children-first and UiText consumes MouseDown (selection),
so the label child swallowed the menu button click and the dropdown never opened.
The transcript similarly gained an invisible Ghosted-button child (a 16x16
selection dead-zone). The old hand-made build never had these — it skipped Type 12
and hand-placed the widgets with no children.
Fix: behavioral widgets (Meter/Menu/Button/Scrollbar/Text/Field) draw their full
appearance and reproduce their dat sub-elements procedurally, so they are LEAF —
the importer must not build their dat children as separate (click-stealing)
widgets. Add UiElement.ConsumesDatChildren (default false; the 6 behavioral
widgets override true) and gate LayoutImporter recursion on it (replacing the
UiMeter-only special case). Only generic containers (UiDatElement, panels) recurse.
Visually confirmed in the live client (channel menu opens; General/Trade selected
and sent). Vitals unchanged (UiMeter was already leaf). Full suite: 404 passed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Record the two execution-time corrections to the design's registration
assumptions: the editable input resolves to Type 12 (Variant B, controller-placed
UiField), and Type 3 is NOT factory-registered (acdream's Type-3 elements are
chrome/containers, kept on the UiDatElement fallback).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Task 6 registered Type 3 -> UiField globally, which broke acdream's Type-3 dat
elements: in these layouts Type 3 is sprite-bearing CHROME (the 8-piece bevel
corners, e.g. vitals 0x10000633 -> sprite 0x060074C3) and the transcript/input
CONTAINER panels — NOT editable fields. UiField draws no dat sprite, so the
vitals bevel corners would render empty; the regression was masked by weakening
VitalsTree_ChromeCornerHasExpectedSprite (UiDatElement+sprite -> UiField+exists).
Retail Type 3 IS UIElement_Field, but retail draws those chrome elements as inert
media-bearing Fields, which our UiDatElement reproduces pixel-for-pixel without a
spurious focus/edit affordance. The one true editable field — the chat input
0x10000016 — resolves to Type 12 and is controller-placed as a UiField (Variant B,
kept). So Type 3 stays on the generic fallback; register it as UiField only when a
window carries a factory-built editable Type-3 field (and UiField grows a
background-media draw + an opt-in editable flag then).
Restored the chrome-corner conformance test (asserts UiDatElement + sprite, an
early warning if Type 3 is ever wrongly routed to UiField). Kept the good Task-6
work: UiField rename + the Variant-B input wiring (stray Type-12 placeholder
removed). Full suite: 404 passed, 2 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Rename UiChatInput → UiField (UIElement_Field, RegisterElementClass(3) @ :126190);
update doc to cite retail's CatchDroppedItem/MouseOverTop drag-drop hooks for
future item windows. BackgroundColor default → transparent (controller sets
the translucent 0.35α value explicitly, matching UiText pattern).
- Register Type 3 in DatWidgetFactory.Create: `3 => new UiField()`.
- ChatWindowController.Bind (Variant B): factory now builds 0x10000016 as an
invisible UiText placeholder (Type 12); Bind removes that placeholder via
FindElement(InputId).Parent.RemoveChild and places a UiField at the same rect.
Result: exactly ONE input widget in the input bar, no stray UiText duplicate.
- Input property type changed from UiChatInput to UiField; GameWindow.cs:1861
UiField.Keyboard assignment compiles unchanged (field exists).
- Tests: UiChatInputTests → UiFieldTests (class + all ctor refs renamed);
DatWidgetFactoryTests: new Type3_Field_MakesUiField test; ChatWindowControllerTests:
updated stale "skipped by factory" comments; LayoutConformanceTests: updated
VitalsTree_ChromeCornerHasExpectedSprite — Type-3 chrome-corner elements are
now UiField (sprite rendering for Type-3 dat image elements is a known
limitation, tracked for post-Task-8 UiField.BackgroundSprite follow-up).
- Full suite: 404 passed, 2 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Rename UiChatView -> UiText (the retail UIElement_Text class,
RegisterElementClass(0xc) @ acclient_2013_pseudo_c.txt:115655).
Factory changes (DatWidgetFactory.cs):
- Remove the Type-12 skip (was: no-media -> null, with-media -> UiDatElement).
- Add Type 12 -> BuildText() -> UiText in the switch.
- BuildText extracts the element's Direct/Normal sprite as BackgroundSprite
so any dat-media the element carried keeps rendering under the text.
UiText changes (renamed from UiChatView.cs):
- BackgroundColor default: (0,0,0,0.35) -> (0,0,0,0) (transparent).
An unbound UiText draws nothing; the controller opts in to the translucent bg.
- New BackgroundSprite + SpriteResolve: optional dat state-sprite background
drawn UNDER DrawFill+text (faithful UIElement_Text media support).
ChatWindowController.cs (Task 5 Step 8):
- Transcript property: UiChatView -> UiText.
- Bind() now uses layout.FindElement(TranscriptId) as UiText (factory-built)
instead of manually constructing + AddChild-ing a new UiChatView.
- Sets BackgroundColor = (0,0,0,0.35) on the found widget (retail translucent bg).
- Removes the tInfo null-check from the early guard (transcript is factory-built;
iInfo lookup kept for the input widget which is still manually constructed).
- BuildLines: UiChatView.Line -> UiText.Line throughout.
Vitals frozen: the Type-12 vitals number elements are meter children and are
never recursed by BuildWidget (the `if (w is not UiMeter)` gate), so they are
not built as widgets and keep rendering via UiMeter.Label. Vitals fixture
vitals_2100006C.json unchanged; LayoutConformanceTests + VitalsBindingTests green.
Tests:
- UiChatViewTests.cs -> UiTextTests.cs (class: UiTextTests, all UiChatView.* -> UiText.*)
- UiChatViewDatFontTests.cs -> UiTextDatFontTests.cs (same)
- DatWidgetFactoryTests: delete Type12_StylePrototype_ReturnsNull +
DatWidgetFactory_Type12WithMedia_Renders; add Type12_Text_MakesUiText +
DatWidgetFactory_Type12_AlwaysMakesUiText.
- LayoutImporterTests: BuildFromInfos_Type12Child_IsSkipped_Type3Present updated
to assert IsType<UiText> (element is now in tree, transparent, not skipped).
Divergence register: AP-37 amended -- removed the "standalone Type-0 text
elements skipped / dat-text widget is Plan 2" clause (now shipped as UiText);
kept the meter-collapse clause and the vitals-numbers-via-UiMeter.Label clause.
AP-38/AP-39/AD-28 file references updated UiChatView.cs -> UiText.cs.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Review caught a behavior divergence: the generic UiMenu auto-set its own
Selected on any enabled pick, while the controller's EnabledProvider keeps the
null-payload specials (Squelch / Tell-to-Selected) enabled/white like retail.
So a special-item click set Selected=null and shifted the highlight onto the
deferred placeholders — and the menu tests masked it by using a different
(specials-disabled) gate than the controller ships.
Fix: clean MVC contract mirroring retail UIElement_Menu::NewSelection — the
widget REPORTS the pick via OnSelect; the controller OWNS Selected (it sets it
only for talk-channel payloads). A special-item click now fires OnSelect(null),
the controller ignores it, and the active channel + highlight stay put —
observably identical to the pre-generalization widget, and extensible for when
Squelch lands. Tests realigned to the controller's gate (specials white) and to
the controller-owns-Selected contract.
Full suite: 403 passed, 2 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
UiChannelMenu → UiMenu: removed ChatChannelKind, the 14-item array, the
button-text map, and the availability default. Generic surface: MenuItem
(label + object? Payload), Selected (object?), OnSelect, EnabledProvider,
ButtonLabelProvider, RowsPerColumn/RowHeight/ColumnWidth (all settable).
All draw/event mechanics unchanged — same popup geometry, same click
coordinates, same 8-piece bevel, same 3-slice button face.
ChatWindowController gains ChannelItems[], ChannelButtonLabel(), and
ChannelAvailable() (verbatim from old widget), and populates the
factory-built Type-6 UiMenu via find-by-id rather than constructing a
replacement widget. The Menu property type is now UiMenu. OnChannelChanged
wrap replaced with the generic OnSelect wrap for the ReflowInputRow hook.
DatWidgetFactory registers Type 6 → new UiMenu().
Tests: UiChannelMenuTests → UiMenuTests (10 tests, all green); factory
Type6 test added; ChatWindowControllerTests updated to use OnSelect.
Divergence register: AP-42 added (flat item model vs retail nested-submenu
MakePopup @0x46d310 — latent, unreachable through the chat menu).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Introduces UiButton: a dedicated dat-widget button that ports UIElement_Button
(RegisterElementClass(1,...) @ acclient_2013_pseudo_c.txt:125828). State selection,
tiled DrawSprite, and label rendering mirror UiDatElement exactly so the chat Send
and Max/Min buttons have zero behavioral change.
DatWidgetFactory now maps Type 1 → UiButton (beside Type 7 → UiMeter, Type 11 →
UiScrollbar). ChatWindowController's Send and Max/Min bind blocks updated from
UiDatElement casts to UiButton casts; ClickThrough=false lines dropped (UiButton
is interactive by construction).
The old UiPanel.cs UiButton (a plain dev-scaffold rect+text button with no dat
sprites) is renamed UiSimpleButton to free the name — no production code
instantiated it.
Full suite: 402 passed, 2 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- git mv UiChatScrollbar.cs → UiScrollbar.cs; rename class + update doc summary to
"Generic scrollbar. Ports retail UIElement_Scrollbar (RegisterElementClass(0xb) @
acclient_2013_pseudo_c.txt:124137); thumb size = trackLen * ThumbRatio (min 8px); step ±1 line."
- git mv UiChatScrollbarTests.cs → UiScrollbarTests.cs; rename test class + replace
every UiChatScrollbar reference with UiScrollbar (bodies unchanged).
- DatWidgetFactory: register Type 11 → new UiScrollbar() before the _ fallback case.
- ChatWindowController: change Scrollbar property type to UiScrollbar; replace the old
"construct-remove-add" block with a "find factory-built UiScrollbar and bind in place"
block (no RemoveChild/AddChild); keep `var track` assignment in scope so the Max/Min
block's track.Left/track.Width reads still compile against UiElement?.
- AP-41 divergence register: update file:line to UiScrollbar.cs:35; narrow wording to
"fallback only — single-tile drawn only when cap ids are unset; the chat controller
passes all three cap ids so the 3-slice path is the active code path."
- Update inline UiChatScrollbar doc-comment references in UiScrollable.cs + UiChatView.cs.
- Full suite: 399 passed, 2 skipped (dat/tower fixture skips), 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Add ChatLayoutFixtureGenerator.cs (Skip-by-default) to regenerate
chat_21000006.json from the live portal.dat via LayoutImporter.ImportInfos
- Commit generated fixture chat_21000006.json (13 KB, 400 lines) — dat-free,
auto-copied to test output via existing *.json csproj glob
- Refactor FixtureLoader: extract shared LoadInfos(fileName) helper; add
LoadChat() + LoadChatInfos() mirroring the vitals pattern; LoadVitalsInfos()
now delegates to the shared loader (behavior unchanged, vitals tests green)
- Add ChatLayoutConformanceTests: ResolvesKnownElements + ResolvedTypes_MatchRetailRegistry
Confirmed resolved Types from live dat:
0x10000011 (transcript) → Type 12 (style-prototype, skipped by factory)
0x10000016 (input) → Type 12 (style-prototype, skipped by factory)
0x10000014 (menu) → Type 6
0x10000012 (scrollbar) → Type 11
0x10000019 (send) → Type 1
0x1000046F (max/min) → Type 1
Also fix pre-existing build break: UiChatInput.MoveCaret(int delta) was made
private in ce848c1 but UiChatInputTests.Backspace_DeletesBeforeCaret called it
as public. Expose a public MoveCaret(int) overload (no-shift) alongside the
private MoveCaret(int,bool) — restores the intended test surface.
Full suite: 398 passed, 2 skipped (generator + pre-existing), 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Capture the 'why beyond chat' the user articulated: chat is the proving ground;
the real payoff is inventory/spell-bar/vendor/character-sheet/trade becoming
data-driven assembly + thin controller. Notes what carries forward (the generic
widget toolkit + the find-by-id controller pattern) vs what those windows still
need (ListBox/Panel + Field drag-drop, the window-manager half of Plan 2, and
per-domain item/container data).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Design for refactoring the hand-named chat widgets + Send/MaxMin click-wiring
into generic, Type-registered widgets built by DatWidgetFactory, collapsing
ChatWindowController (and, gated-last, VitalsController) to a thin retail
gm*UI::PostInit-style find-by-id binder.
Key finding that reframes the pass: the importer's base-chain Type resolution
is already retail-faithful, and Type 12 is UIElement_Text (a real behavioral
class), not a style prototype to skip — verified against
acclient_2013_pseudo_c.txt:115655. The generalization is therefore a
registration task (register Types 1/3/6/11/12 -> generic widgets, delete the
Type-12 skip), not a new mechanism.
Approved scope: full registry (bounded to the Types chat+vitals use; rest stays
UiDatElement fallback), chat-first, vitals rewire as the final separately-gated
step. 7-step one-widget-per-commit migration; new chat_21000006.json golden
fixture; vitals fixture stays frozen through steps 1-6.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A `dump-font-atlas` subcommand renders a dat Font's fg/bg atlases (alpha as
luminance) plus a sample string composited exactly the way DrawStringDat does
it (outline + fill, integer-snapped). Used to reproduce the glyph-baseline
jitter offline (fractional-origin bug vs the fix) without launching the client.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- ChatWindowController: wires the menu chrome (popup bevel, row/checkbox
sprites), the input focused-field sprite + keyboard, and autosizes the channel
button + reflows the input field to start after it (anchor re-capture so the
per-frame layout doesn't fight it). DefaultTextInput / write-mode focus hooked
up.
- WrapText now breaks an over-long UNBROKEN token at character boundaries (no
hyphen), packed onto the current line first — so a spaceless token wraps
instead of overflowing, and a "You say," prefix stays on the same row as the
start of the message.
- UiChatView: transcript background + selection highlight use DrawFill (sprite
bucket) so the transcript text draws ON TOP instead of being dimmed by its own
translucent rect background.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
In chat write mode the keyboard belongs to the input — typing "swd" must not
walk the character — but AUTORUN must keep going (the user can chat while
running).
- InputDispatcher.IsActionHeld now returns false while WantCaptureKeyboard is
set (a focused chat input), the polling-path twin of the existing gate on
Fired actions. This SUPERSEDES the old per-frame OnUpdate early-return, which
also killed autorun. Gating here instead lets the movement block keep running,
so autorun — a separate latched bool ORed into Forward at the call site, not a
polled key — survives. Test updated to encode the new contract.
- GameWindow: the movement suppress-guard reverts to ImGui-devtools-only (the
retail write mode no longer early-returns); wires DefaultTextInput = the chat
input (Tab/Enter activation) and Input.Keyboard for clipboard. Drops the
one-shot UI-scale diagnostic.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Port the retail UIElement_Text editable single-line field:
- Focused = "write mode": draws the gold lit field sprite (0x060011AB, the
Normal_focussed state) instead of the flat translucent rect; Enter submits
AND blurs (exits write mode).
- Single-line SELECTION: click-drag, Shift+Arrows, Shift+Home/End, Ctrl+A;
translucent-blue highlight behind the span; typing/Backspace/Delete/Paste
replace the selection first.
- CLIPBOARD: Ctrl+C copy, Ctrl+X cut, Ctrl+V paste at the caret (control chars
stripped — single-line). Wired to the keyboard device for clipboard + Ctrl/
Shift state.
- Held-key AUTO-REPEAT (Silk delivers one KeyDown per press): Backspace /
Delete / Left / Right repeat via a 0.4s-delay, ~25/s OnTick timer.
- Horizontal SCROLL + clip: keeps the caret in the field and draws only the
glyph window that fits inside it, so long input scrolls within the box
instead of spilling past Send into the 3D world.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Match the talk-focus menu + button to retail (decomp-verified):
- Menu item text is FILL-ONLY (retail UIElement_Text outlines only when
SetOutline(true); the talk-focus items don't) — kills the grey halo. Available
items render white; UNAVAILABLE items render grey (not the salmon colorPink,
which is a chat-MESSAGE color we'd misapplied). Special items (Squelch /
Tell-to-Selected) render white. Labels indent past the baked checkbox in the
row sprite (0600124E empty box / 0600124D white checkmark) instead of
overlapping it.
- The popup is wrapped in the universal 8-piece window bevel (the menu sprite
family has no border) and draws in OnDrawOverlay so the translucent chat
panel no longer greys its right column.
- The button face (0600124D/66, a fixed 46px LED+arrow sprite) is now 3-sliced
(LED cap / stretch / arrow cap) and autosizes to its label via
NaturalButtonWidth, so "Chat" fits in the body instead of running into the
arrow. The status LED (red Normal / green Pressed) is no longer overdrawn.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The retail-look render + focus primitives this chat pass builds on:
- TextRenderer: an OVERLAY layer (sprite/rect/text buckets flushed AFTER the
normal layer) so an open popup composites on top of everything incl. rect
panel backgrounds; a DrawFill primitive (solid quad via a 1x1 white texture)
routed through the SPRITE bucket so a panel background draws UNDER its text
instead of being washed by the later rect bucket; and the text pass now
disables SampleAlphaToCoverage + Multisample so glyph alpha edges aren't
dithered into MSAA coverage (the "fuzzy text") — self-contained GL state
per feedback_render_self_contained_gl_state.
- UiRenderContext.DrawStringDat: snap the line baseline to a whole pixel ONCE
then add the integer per-glyph offset (retail DrawCharacter takes an int
pen-Y + schar m_VerticalOffsetBefore) — fixes the "letters dip down" jitter
at a fractional line origin. Outline pass is now opt-in (retail gates it per
element via SetOutline; default off = crisp fill-only). Adds DrawFill +
Begin/EndOverlayLayer.
- UiElement: OnDrawOverlay + DrawOverlays (second traversal), FindRoot (blur
self), ResetAnchorCapture (re-baseline an anchored element after reflow).
- UiRoot: runs the overlay pass after the main tree; Tab/Enter focuses the
DefaultTextInput (write-mode activation); a left click on a non-edit target
blurs the focused input (exit write mode without submitting).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
fix(D.2b): point-sample the dat-font atlas so UI text is pixel-crisp
The font glyph atlas was uploaded with bilinear (Linear) min/mag filtering, which
softens the small dat-font glyphs (the menu/button text "blur"). Add a nearest-filter
path to UploadRgba8/GetOrUploadRenderSurface and use it for the font atlases only
(world + other UI surfaces keep bilinear). Combined with the existing per-glyph
pixel-snap, glyph texels now map 1:1 to screen pixels. Sharpens all dat-font text
(transcript, menu, Send/Chat buttons, vitals numbers).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
fix(D.2b): arrow swap, centered menu text, scrollbar-to-top, Send caption
- scroll arrows: native sprites are opposite (0x06004C6C up / 0x06004C69 down) per live
visual — swap the assignment, drop the V-flip.
- menu labels centered vertically in each 17px row (was top-aligned, looked corrupt).
- scrollbar pulled up to the panel top so the top arrow meets the window border and the
max/min button lines up with it (the 6px dat offset left a gap after the resize-bar reclaim).
- Send button: the dat sprite 0x06001915 is a blank gold frame (export-confirmed), so add a
generic optional Label/LabelFont to UiDatElement and draw "Send" centered on it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
feat(D.2b): data-driven channel menu chrome + greying + scroll-arrow fix
Investigation found the menu popup is fully dat-driven (UIElement_Menu::MakePopup
@0x46d310 reads LayoutDesc 0x21000006 elements 0x1000001C/1D/1E — the "stray" top-level
elements). Render the popup from the real sprites instead of a flat rect:
- panel 0x0600124C, item row 0x0600124E, selected row 0x0600124D; 191x17 rows, 2 cols.
- drawing rows as SPRITES also fixes the z-order (a DrawRect bg composited OVER the
labels; sprites share the labels submission bucket so text lands on top).
- item greying: available channels white, unavailable salmon (colorPink) — static
approximation (Say/General/Trade/LFG) with an AvailabilityProvider hook for live
TurbineChat state; unavailable items are inert on click. Ports ResetAllTalkFocusMenuButtons.
- scroll arrows: both dat sprites point down (export-confirmed); V-flip the top button
so it points up.
Tabs confirmed to have NO digits in retail (blank gold frames) — acdream already matches.
Build + 392 App tests green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
fix(D.2b): channel menu popup opaque + button label tracks selected target
- the popup inherited the chat window 0.75 opacity so the transcript bled through;
add UiRenderContext.PushAlphaAbsolute and draw the popup at absolute opacity.
- the "Chat" button was hardcoded; it now shows the active talk target (retail
updates it on selection). Exact textured menu-panel sprite is a follow-up (the
popup is a keystone UIElement_Menu construct, not in the chat LayoutDesc).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
feat(D.2b): exact retail chat colors from a live cdb dump
Attached cdb to a live retail acclient (PDB-matched) and read the named RGBAColor
constants at acclient 0x81c4a8+ (colorWhite/colorBrightPurple/colorLightBlue/
colorGreen/colorLightRed/colorGrey), used by ChatInterface::BuildChatColorLookupTable
@0x4f31c0. Replaced the approximated RetailChatColor palette with the ground-truth
values: speech=white, tell=colorBrightPurple(1,.498,1), channel=colorLightBlue
(.247,.749,1), system/popup=colorGreen(.5,1,.498), combat=colorLightRed, emote=colorGrey.
Capture scripts saved under tools/cdb/.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
feat(D.2b): chat polish — typing fix, opacity, scrollbar 3-slice, retail channel menu
Visual-iteration batch (decomp-grounded), each fix verified against the retail screenshots:
- typing: UiElement.HitTest aborted on ClickThrough BEFORE walking children, so the
ClickThrough UiDatElement panels blocked hit-testing to the input/transcript inside
them. Check ClickThrough AFTER the child walk (it only gates whether THIS element
claims the hit). Restores input focus + typing.
- opacity: UiElement.Opacity + a UiRenderContext alpha stack applied to sprite/rect
draws (text bypasses it, stays sharp); chat frame Opacity=0.75 → translucent chat.
- brown sliver: grow the transcript panel up 9px to cover the dropped resize-bar strip.
- scrollbar: real 3-slice thumb (caps 0x06004C60/66 + tiled mid) + tiled track.
- max/min: shifted one button-width left of the scrollbar (dat right-anchors collide).
- system text now green (retail ChatMessageType 5; was yellow).
- word-wrap: transcript lines wrap to the panel width (greedy, ports GlyphList::Recalculate).
- channel menu reworked to retail gmMainChatUI::InitTalkFocusMenu: "Chat" button + a
TWO-COLUMN popup of the 14 talk-focus items (Squelch, Tell to Selected, Chat to All,
Tell to Fellows, ...) on a tan panel; channel items set the active outbound channel.
Build + 392 App tests green. Visual confirmation in progress.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@
The live session + its LiveCommandBus are created after the retail-UI block in
OnLoad, so binding the bus by value captured NullCommandBus and silently dropped
outbound chat. Pass a Func<ICommandBus> resolved at submit time (mirrors how the
ImGui ChatPanel re-reads the bus each frame).
AP-41: scrollbar thumb drawn as single stretched tile (0x06004C63) instead of
retail's 3-slice top-cap/middle/bottom-cap — acknowledged in UiChatScrollbar.cs:37,
registered per the divergence-register rule.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the hand-authored chat block (UiNineSlicePanel + inline UiChatView
+ local BuildRetailChatLines/RetailChatColor statics) with
ChatWindowController.Bind(LayoutDesc 0x21000006) — the same LayoutImporter
path as the vitals window. The controller places UiChatView (transcript) +
UiChatInput (text entry, on-submit) + UiChatScrollbar + UiChannelMenu inside
the dat-authored chrome. The dead local statics are deleted.
Wired to _commandBus (same LiveCommandBus as the ImGui ChatPanel) so
type+Enter dispatches SendChatCmd server-ward. Transcript keyboard set from
_uiHost.Keyboard (set by WireKeyboard above the chat block) for Ctrl+C/Ctrl+A.
Divergence register: added AD-28 (two-widget split vs UIElement_Text),
AP-38 (no in-element word-wrap), AP-39 (per-line colour vs per-glyph runs),
AP-40 (no opacity fade / shared vitalsDatFont), TS-30 (tab buttons no-op),
TS-31 (no squelch); updated IA-15 to cover both vitals + chat importer paths.
Build: 0 errors/warnings. Tests: 392 passed, 1 skipped (expected).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Implements Task G2: binds the imported chat LayoutDesc (0x21000006) to live
behavior, the acdream analogue of retail ChatInterface + gmMainChatUI::PostInit.
- UiDatElement: add OnClick hook + OnEvent override so Send/max-min buttons
can be wired by a controller without needing a dedicated widget type.
- ChatWindowController.Bind: reads transcript (0x10000011) and input
(0x10000016) rects from the raw ElementInfo tree (factory skips them as
Type-12/no-media), places UiChatView under the transcript panel and
UiChatInput under the input bar; replaces the imported scrollbar track
(0x10000012) with UiChatScrollbar driving UiChatView.Scroll; replaces
the channel menu placeholder (0x10000014) with UiChannelMenu; wires
Send button and max/min toggle via the new OnClick hook.
ChatCommandRouter.Submit routes all input through the existing pipeline.
- 6 smoke tests: Bind returns non-null, Transcript is child of panel,
Input is child of bar, Input.OnSubmit publishes SendChatCmd, channel
change updates submit channel, returns null when panels missing.
Build: 0 errors. Test suite: 392 passed / 1 skipped / 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Task G1: two gaps blocked chat window static sprite elements from rendering.
Change 1 — DatWidgetFactory: only skip Type-12 elements that have no own
state media (pure style prototypes). A Type-12 element that carries sprites
(e.g. a chat Send button whose derived Type-0 element inherited Type 12 from
its base prototype) now renders as a UiDatElement.
Change 2 — ElementInfo: add DefaultStateName field (string, default "").
Change 3 — LayoutImporter.ToInfo: read ElementDesc.DefaultState.ToString()
into DefaultStateName; normalize Undef/Undefined/0 sentinels to "".
Change 4 — ElementReader.Merge: inherit DefaultStateName (derived wins if
non-empty, else base).
Change 5 — UiDatElement ctor: initialize ActiveState to DefaultStateName
when set; else "Normal" when a Normal-state sprite is present (retail's
implicit default for buttons/tabs); else "" (DirectState). This makes the
Send button, max/min button, and numbered tabs render their default sprite
without requiring explicit state assignment at runtime.
Vitals neutrality: all vitals chrome/grip elements carry DirectState-only
sprites with no "Normal" named state and DefaultStateName="" (Undef in dat),
so their ActiveState stays "" and their existing conformance tests are
unaffected. Vitals text labels (Type 0→12 via Merge, no StateMedia) are
still skipped by the refined Type-12 guard (StateMedia.Count==0).
Tests: 4 new tests (2 in DatWidgetFactoryTests, 3 in UiDatElementTests).
All 386 pass; 387 total (1 pre-existing skip).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Outdoor objects brightened as the camera approached: lighting selected the
nearest 8 lights to the VIEWER and fed that one global set to everything
(LightManager.Tick), so a building's wall torches only lit it once the camera
got close enough for them to win the global top-8. Probe confirmed the scale of
the problem: a single Holtburg view registers 129 point lights — the global cap
of 8 was hopeless.
Retail selects up to 8 lights PER OBJECT by the object's own position
(minimize_object_lighting 0x0054d480), so a torch always lights the wall it
sits on, camera-independent. Ported faithfully:
- LightManager.SelectForObject (pure, TDD, 8 new tests): candidacy
(light.pos − center)² < (Range + radius)², nearest-8 among those. Plus
BuildPointLightSnapshot for the per-frame stable-indexed light list.
- mesh_modern.vert: two SSBOs — binding=4 GLOBAL point-light array (the
snapshot), binding=5 per-instance light SET (8 int indices into it, -1 =
unused), parallel to the binding=0 instance buffer (mirrors the U.3 clip-slot
mechanism). accumulateLights keeps ambient + sun from the SceneLighting UBO
(cleared as faithful by the lighting audit) and loops THIS instance's point
lights. pointContribution factored out (same calc_point_light wrap+norm shape).
- WbDrawDispatcher: per-entity light set computed ONCE at the isNewEntity site
(constant across the entity's parts), by the entity's AABB sphere; threaded
into grp.LightSets parallel to grp.Matrices; global + per-instance buffers
uploaded in Phase 5. Camera-independent ⇒ stable for static buildings.
- GameWindow: BuildPointLightSnapshot + dispatcher.SetSceneLights each frame.
Tests: 17/17 LightManager + 36/36 dispatcher clip-slot/clip-frame green
(parallel-array lockstep preserved). Visually gated: the meeting hall now holds
steady as the camera approaches (was the popping symptom).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
