The interactive 2-node playtest wave -- every fix decomp-grounded and live-verified:
COLLISION ECONOMY (the ram one-shot): StaticBounce mutates worldLinearVelocity
per contact and ProcessCollisionList walks EVERY touched solid per frame; with
2007 terrain-as-solids the reflections compounded x4-x40 within one frame and a
walking bump one-shot a pristine mech for 112,375 pts (62-pt authentic economy).
Fix: frameEntryWorldVelocity restore per contact (damage always priced at the
real approach speed -- all the binary's physics ever saw); Mech::Reset zeroes
the mover motion (respawn = teleport); [collide-tx]/[mp-hdlr] telemetry.
Gotcha #16 (engine-facility drift class).
MISSILES: peer-visible salvos (the launcher record extension carries a salvo
counter + aim point; ForceUpdate actually enqueues it -- the dirty flag alone
never serialized), the authentic arc (authored MuzzleVelocity vector + the
Seeker's 200m/0.1/300 loft + gain-4 steering, decoded from @004beae4/@004bef78),
world-impact bursts (rounds detonate on cave geometry instead of phasing
through), contact-only damage (flight-cap expiry = fizzle, no more teleport
damage), live re-lead, and ballistic (unguided) shells for autocannons.
projweap's stale BTPushProjectile extern (the /FORCE signature trap, gotcha #6
corollary) crashed the Avatar's first AFC100 shot -- fixed + sweep rule recorded.
RADAR: two transcription bugs made the scope permanently empty -- FUN_0040b244
is the affine INVERSE (not a copy) and FUN_0040adec writes ONLY the 3x3 rotation
(never the translation); worldToView now Invert(view) built rotation-first.
CulturalIcons sorted out of the moving grid (the phantom red pips were map
props), visible-radius culls on all three draw passes, live pip verified at
|delta| x ppm px. Gotcha #17 (verify the FUN_ body, not its call shape).
WEAPON PANELS (the frozen-dial hunt): the binary's *(subsystem+0x40) means
FAILED -- the recon's 'operating' name was backwards, inverting the destroyed-X
lamps, the panel look, the children enable and the ready-lamp gate (which had
NEVER executed). Polarity chain corrected end-to-end (failedState, fed by real
damage saturation). Root cause of the freezes: MFD page-mode gating -- the dev
composite shows ALL pages at once, so off-page dials legitimately stopped; under
BT_DEV_GAUGES the 15 page-plane bits stay active (the exclusive secondary trio
untouched). The SEH gauge guard now names its kills; repaint-heal resets the
incremental arc after panel repaints; [panel]/[arc] probes added.
COMM/SCORE: MessageBoard LIVE (the engine already shipped the whole
Player__StatusMessage queue; wired the binary's one producer -- the kill branch,
victim's name, 6s -- plus the consumer bridge and a lazy source bind); MP DEATHS
counted via the observed-death tally (each node scores every pilot from locally
observed events, the same model as the KILLS credit) and the -2/-1 engine seed
clamped for display.
DEV UX: node-tagged window titles (-net port), gauge panel reworked (1320x480,
true 4:3 MFD cells, the portrait secondary UNROTATED upright, linear filtering,
BT_GAUGE_SCALE), fixed close spawns via BT_SPAWN_XZ, Boreas flies an Avatar
(first second-chassis live outing).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
TASK #17 -- AUTHENTIC RADAR SYMBOLOGY (the cross-blip stand-in retired).
The recon proved the 'missing pip raster set' never existed as BT game code:
pips are the ENGINE's L4GaugeImage vector-stroke system (T0 source in tree,
L4GAUIMA.cpp == FUN_0046f0c0 line-for-line), the 'pip table' is L4Warehouse::
gaugeImageBin keyed by Entity::resourceID, and BTL4.RES ships 110 type-0x12
shapes (every mech/vehicle/building/tree). The six btl4rdr stubs are wired
to the real facilities: contacts draw their authentic model silhouettes with
LOD selection, the target gets the binary's 4px-inflated box highlight, and
player-name labels resolve via Mission::GetSmallNameBitmap (the prebuilt
64x16 egg rasters keyed by the player's bitmapindex). DECODED: the invented
'VideoObject' was Entity::owningPlayer all along (+0x190; nameID =
playerBitmapIndex@0x1E0, target = BTPlayer::objectiveMech@0x284 -- new
bridge BTPlayerObjectiveMechOf); the 'LabelledEntity' class is Landmark
(cultural.h; label path dormant -- no landmark content ships, no runtime
landmarkID writer). The L4GREND BT_DEV_GAUGES warehouse guard is removed
(its AV had a different culprit, below); resource type 18 corrected to
GaugeImageStream in decomp-reference (was mislabeled 'ModelList').
TASK #14 -- the Myomers ODR duplicate ELIMINATED: the powersub.cpp/hpp
'Myomers' (classID 0xBC3 -- actually Sensor) is retired whole; it duplicated
?DefaultData@Myomers@@ against the real class (dumpbin-verified) and /FORCE
picked the winner by link order. The real Myomers (0xBC6) now chains
PoweredSubsystem's handler set (ids 4-8) and publishes its SEVEN binary
attributes (@00511588: SpeedEffect/Current/Recommended/Min/MaxSeekVoltage-
Index/SeekVoltage/OutputVoltage) -- the old empty unchained index starved
the Myomer engineering panel of every resolve.
TASK #16 -- attribute parity: MechWeapon publishes the FULL binary table
@0x511890 (11 entries; ids renumbered to binary truth -- the port aliases
had squatted the binary's DistanceToTarget/TargetWithinRange ids; the
streamed TriggerState 0x13 binding unchanged). Binary names resolved two
TODO members: pipState -> estimatedReadyTime (attr 0x1A), and the EXT-model
flag is the binary's RearFiring (0x1B). ThermalSight LightState published.
HUD (offset conflicts) + missile-side tables (id encoding suspect)
documented for a re-dump instead of publishing blind.
THE CRASH THIS EXPOSED [T2, cdb-verified]: gotcha #11's dense-table gap
is a LATENT AV, not a guaranteed one -- the old table's 0x0D..0x12 gap
(task #5) survived on heap luck; the renumber reshuffled allocations and
Find() AV'd on a garbage entryName in WeaponCluster's PercentDone resolve.
Fixed with five named PAD entries (the mech.cpp attrPad idiom) + a
static_assert locking the pad base to PoweredSubsystem::NextAttributeID.
Gotcha #11 amended with the proof.
TASK #15 -- stale-ledger sweep: GAUGE_COMPOSITE ('composite not yet built',
Reservoir shadow, PlayerStatus/vehicleSubSystems 'remaining', valve-dormant-
until-0xBD3, sensor guard, the superseded 'Heat MFD near-static' reframe --
all banner-corrected), gauges-hud frontmatter, L4VB16 + powersub comments.
Verified live: 50/50 config attribute bindings resolve, 0 NULLs, 0 parse
skips, mech spawns and simulates 31/31 subsystems, no cross-blip fallbacks,
the pip cache fills through entity registration without the old guard.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Mark PlayerStatus reconstructed + registered in the remaining-widget scope note;
record the key finding that it lives in cameraInit (the mission-review camera
cockpit) not MechInit, so it is build/link/mech-cockpit-safe but not runtime-
verifiable via a normal mech egg. vehicleSubSystems (from-scratch, 26 uses)
remains.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Both need a full multi-function reconstruction (not the registration-glue the
other widgets were): PlayerStatus's ctor/Execute/helpers (CreateMutantPixelmap8,
PlayerStatusMappingGroup, App Players-node lookup) are comment-only stubs and its
existing Make/header ctor are buggy (drop the graphics_port_number param, pass
entity/renderer as x/y); vehicleSubSystems is undeclared/from-scratch. Records
the decomp addresses + the ctor arg order for a focused follow-up.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Update GAUGE_COMPOSITE.md + CLAUDE.md: the radar plots the enemy contact inside
the FOV wedge. Record the two fixes (engine RebuildEntityGrid populating the
dropped entity feed; the delta-relative blip projection replacing the wrong
camera-matrix point transform) and narrow the remaining map work to the authentic
pip/name infrastructure + SetTargetRange zoom.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
GAUGE_COMPOSITE.md + CLAUDE.md: add the map/MapDisplay radar (increment 6, the
view wedge renders) to the done list; note the systemic pattern (a newly-consumed
reconstructed gauge is the first reader of stubbed helpers -- the map exposed 3);
narrow the remaining gauge widgets to PlayerStatus + vehicleSubSystems, and list
the map follow-ups (contact classification / stubbed pip+name infra, SetTargetRange
zoom).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
GAUGE_COMPOSITE.md + CLAUDE.md: add the OneOfSeveral/oneOfSeveralPixInt button-
lamp reconstruction to the done list; narrow the remaining widgets to map (radar
marquee), PlayerStatus, vehicleSubSystems, plus the follow-ups to make the
duck/searchlight buttons dynamic (extend the Mech table to 0x37 + a Searchlight
table).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Document the cockpit-gauge attribute-pointer system in GAUGE_COMPOSITE.md
(§Attribute wave) and CLAUDE.md §10: how a Subsystem/Attribute or bare-Attribute
binding resolves (ParseAttribute -> FindSubsystem/GetName -> GetAttributePointer
-> activeAttributeIndex), what a class must publish (AttributeID enum +
AttributePointers[] + GetAttributeIndex chained + DefaultData wiring), the
DENSE-TABLE HAZARD (Build leaves gap slots uninitialized, Find strcmps every
slot -> tables must be a dense prefix), and the base-primitive vs BT-specific
widget split. Records the 4 committed increments (HeatSink table, Mech table,
vertBar, segmentArcRatio) and the follow-ups (oneOfSeveralPixInt/map/PlayerStatus/
vehicleSubSystems, the #if0'd HeatSink-bank AmbientTemperature, the Reservoir
coolantCapacity shadow).
Promote the two reverse-engineering helpers from the session scratchpad into
tools/ (durable): disas2.py (capstone + PE parse -- recovers x87 float math
Ghidra drops; FUN_004dcd94=round, FUN_004dcd00=fabs) and vtdump.py (dump a class
vtable to find an override the assert-anchored decomp never exported, e.g.
SegmentArcRatio::Execute). Doc references updated scratchpad/ -> tools/.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Reconstructs ColorMapperCritical -- tints a schematic colour by a subsystem's
operational state (the "critical" secondary display mode):
- CriticalConnection (@004c3598/@004c3610): src==0 -> 0; src->simulationState==1
(DestroyedState) -> 100; else the subsystem's own damage-zone damageLevel*100.
- ColorMapperCritical::Make/ctor (@004c3ddc/@004c3e40): FindSubsystem by name +
wire the CriticalConnection.
Resolved the subsystem damageZone@0xE0 shadow: MechSubsystem::damageZone is
declared ReconDamageZone*, but the assignment (mechsub.cpp: damageZone =
(ReconDamageZone*)new DamageZone(...)) shows the pointer IS a real engine
DamageZone -- so cast it back and read damageLevel. Added public accessors
GetSimulationState()/GetDamageZoneProxy() to MechSubsystem (those fields are
protected; mirrors how HeatConnection reads the public currentTemperature).
Completes the ColorMapper family (cmHeat, cmArmor, colorMapperMultiArmor,
cmCrit). Verified: parses, builds, all subsystems resolve (0 warnings), no
/FORCE unresolved, combat un-regressed (TARGET DESTROYED, 0 crashes), stable.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Reconstructs colorMapperMultiArmor -- tints one schematic colour (a whole torso
section) by the WORST of up to 8 damage zones:
- MultiArmorConnection (@004c346c/@004c34f4): scans 8 zones via
entity->damageZones[idx], keeps the max damageLevel, count==0 ? 100 :
Round(worst*100).
- ColorMapperMultiArmor::Make/ctor (@004c3c48/@004c3d60): resolves 8 named
zones to indices via Entity::GetDamageZoneIndex (13-param methodDescription).
ALSO fixes a latent HEAP CORRUPTION that registering this widget exposed (and
that would have blocked every further widget): GaugeInterpreter's bytecode
buffer is a fixed interpreterTableSize=46864 (GAUGREND.h) tuned for RP's config,
and its Insert() bounds guard is a Verify() that compiles out at DEBUG_LEVEL 0.
As each BT gauge widget is registered, more of BT's larger L4GAUGE.CFG resolves
into bytecode (vs being parse-skipped); colorMapperMultiArmor's 13 params x
hundreds of calls pushed the total past 46864 -> silent overflow past the char[]
-> heap smash detected later in mission bitmap loading (not the gauge code).
Fix: interpreterTableSize -> 262144 (sized for BT's full config).
Verified: parses, builds, all zones resolve, no /FORCE unresolved, combat
un-regressed (TARGET DESTROYED, 0 crashes), stable. Details: docs/GAUGE_COMPOSITE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Each armor zone on the cockpit schematic is now tinted by that zone's live
damage. Render-verified: the schematic shows an all-green Blackhawk (all zones
damageLevel=0 at spawn), replacing the static green+red it showed before cmArmor
was registered -- i.e. cmArmor now owns those zone colors and shows the real
undamaged state; damaged zones shift toward red.
Reuses the ColorMapper base (from cmHeat) + two pieces reconstructed from the
binary (Ghidra dropped the x87, recovered by disassembly):
- ArmorZoneConnection (@004c33a4/@004c3430): resolves one zone from the owner's
inherited Entity::damageZones[zone_index]; per-frame feed = zone==NULL ? 100
: Round(zone->damageLevel * 100) (the 0..1 damage ratio -> 0..100 percentage
-> the colour index ColorMapper::Execute pushes into the palette slot).
- ColorMapperArmor::Make/ctor (@004c3aa4/@004c3b98): Make resolves the CFG zone
name (dz_ltorso etc., the 6th param) to an index via Entity::GetDamageZoneIndex
(== the binary's FUN_0042076c, scanning damageZones[] by name); the ctor wires
the ArmorZoneConnection.
Decomp fact (corrects the cmHeat note): the ColorMapper base ctor takes NINE
args -- an owner_ID sits between renderer and graphics_port_number. Our 8-arg
ColorMapper::ColorMapper folds owner_ID=0 in (gauges have owner 0), so it is
equivalent; cmArmor's mapping matches cmHeat plus the zone name.
Verified: parses, builds, all dz_* zones resolve (0 "not found"), no /FORCE
unresolved, combat un-regressed (TARGET DESTROYED, 0 crashes). The dynamic
red-on-damage needs the player to take damage (the passive spawn dummy never
hits back). Details: docs/GAUGE_COMPOSITE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The visible money-shot: a green compass needle that rotates with the mech's
facing plus a numeric heading in whole degrees (render-verified: the needle
swung and the number went 247 -> 182 as the mech turned).
Full class reconstruction from the binary (Make/ctor/dtor/TestInstance/
ShowInstance/BecameActive/Execute). Ghidra dropped the x87 float math feeding
Execute's needle endpoints and the ctor's NumericDisplay centering, so it was
recovered by disassembling BTL4OPT.EXE with capstone (scratchpad/disas_hp.py):
the needle is a radial line from innerRadius(20) to outerRadius(39) at the
heading angle, endpoints rounded half-up; the readout is round(360 - deg).
Two corrections vs the map:
- The header ctor was 12 args; the binary's is 14 -- widened it. Fields @0x98/
@0x9C are the needle's inner/outer RADIUS (Execute multiplies them by sin/cos),
not "color/spacing" as the old field names implied; renamed accordingly.
- ENGINE-CONVENTION FIX: the binary read the heading from EulerAngles index [0],
but the WinTesla EulerAngles(Quaternion) decomposition is ambiguous for a
yawing mech (the quaternion double-cover flips it to a pitch=roll=pi branch as
yaw sweeps past +/-pi -> the needle spins erratically). Switched to
YawPitchRoll, whose yaw-first .yaw is the clean continuous heading
(pitch=roll~0). Faithful to the binary's intent (heading), correct for this
engine's decomposition.
renderer->GetLinkedEntity() resolves the viewpoint mech (a GetViewpointEntity
fallback is kept belt-and-braces). Deps (NumericDisplay/GraphicsViewRecord/
GraphicGauge) are real engine classes -> no /FORCE risk; verified no unresolved
HeadingPointer externals, no parse desync, combat un-regressed (TARGET
DESTROYED, 0 crashes). Details + the disasm technique: docs/GAUGE_COMPOSITE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Begins the BT gauge WIDGET reconstruction -- making the cockpit instruments show
live data instead of just frame art. The BT gauge method table
(BTL4MethodDescription[]) was a chain-only stub with zero registered
methodDescriptions, so every BT-specific gauge keyword was parse-skipped.
Registers the first widget, cmHeat (ColorMapperHeat -- tints the heat/critical
schematic palette by a named subsystem's live temperature):
- Added ColorMapperHeat::methodDescription (keyword "cmHeat", param list
R,Md,C,St,St,St matching the binary .data) and registered it in
BTL4MethodDescription[] (chain link stays last).
- Rewired ColorMapperHeat::Make to read methodDescription.parameterList[]
(the interpreter restores each instance's parsed params there) instead of the
earlier unrecovered DAT_* placeholders -- so it uses the real CFG rate/mode/
colour-slot/palette/subsystem-name, and the runtime port arg as the graphics
port.
Registering it exposed the /FORCE trap: two REAL functions in the chain were
undefined (silently stubbed -> would AV once the gauge builds). Reconstructed
both from the decomp:
- ColorMapper::BecameActive (@004c395c): invalidate the cached colour index +
last-written RGB so the next Execute re-pushes the hardware palette.
- ColorMapperHeat::~ColorMapperHeat: empty -- the base chain (~ColorMapper ->
~Gauge) releases the HeatConnection + palettes.
Verified: register->parse->Make->ctor->FindSubsystem("GeneratorA") (resolves,
no "does not exist")->HeatConnection->Execute->palette-push runs end-to-end
under BT_DEV_GAUGES; no parse desync; combat un-regressed (TARGET DESTROYED,
0 crashes). Establishes the registration recipe + the /FORCE-check discipline
for every remaining widget. Details + priority order: docs/GAUGE_COMPOSITE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Generalizes the single-'sec'-inset (Milestone B) into a full 6-surface
compositor presented in its own top-level window -- the default under
BT_DEV_GAUGES (BT_DEV_GAUGES_DOCK=1 docks it into the main window instead).
A 960x384 window tiles all six pod instrument screens: Heat (COOLANT/BALANCE/
RES + condenser gauges), Comm (KILLS/DEATHS/SELECT TARGET), Mfd1/2/3 (DISPLAY/
PROGRAM/NEAREST frames), and the color radar (SCALE grid + dials + ARMOR
DAMAGE). Main 800x600 3D view is un-occluded; default DEV un-regressed
(TARGET DESTROYED after 8 hits, 0 crashes).
Design (mapped by the mfd-multisurface-map workflow):
- All 6 surfaces are bit-plane MASKS over the ONE shared SVGA16/pixelBuffer
(sec=palette low byte; Heat=0x4000 UL, Mfd2=0x0400 UC, Comm=0x8000 UR,
Mfd1=0x0100 LL, Mfd3=0x1000 LR), so the compositor reaches the SVGA16 once
and extracts each by mask. No port-name reconcile needed on the dev path --
fetch the BT names directly; the RP aux* names only matter to the pod's own
SVGA16::Update demux (a deferred pod-only fallback).
- SVGA16::DrawDevSurface: two kernels -- palette-LUT (sec) + mono bit-plane->
tint ((word&mask)?tint:0). BTDrawGaugeSurfaces iterates a 6-entry table.
- Separate window = one CreateAdditionalSwapChain on the existing device (no
2nd D3D device). BTGaugeWindowRenderAndPresent (after the main EndScene):
SetRenderTarget -> SetDepthStencilSurface(NULL) -> Clear -> BeginScene ->
6 tiles -> EndScene -> restore -> swap->Present.
- KEY BUG fixed: the main 800x600 depth surface stays bound when rendering to
the 960x384 gauge backbuffer; a bound depth smaller than the RT is invalid
-> all draws silently fail (clear color, no geometry). Unbind depth (Z is
off), restore after.
Pod SVGA16::Update/BuildWindows path byte-unchanged; all gated BT_DEV_GAUGES.
Content is still the authored cockpit frames + base-table gauges -- the live
MFD widgets need the BTL4MethodDescription reconstruction (this window is now
the live viewer for that work). Details: docs/GAUGE_COMPOSITE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The pod's secondary/radar cockpit surface now renders as an inset in the
800x600 dev window under BT_DEV_GAUGES: radar/tactical grid, SPEED/HEADING/
MAGNETIC/PROP dials, and the color-coded ARMOR DAMAGE mech schematic -- real
gauge content (nzSec=27247), composited from the shared CPU pixelBuffer.
Composite pass (engine): SVGA16::DrawDevInset palette-expands the secondary
plane into a MANAGED texture on the MAIN device and draws an XYZRHW inset quad;
BTDrawGaugeInset() reaches the gauge renderer's 'sec' port and is called from
DPLRenderer::ExecuteImplementation as the last draw before EndScene.
Three reconstruction bugs fixed to get real content:
1. BTL4Application::MakeGaugeRenderer signature mismatch (REAL fix): the
reconstructed no-arg override only HID the 2007-engine's widened 3-arg
virtual MakeGaugeRenderer(int*,int*,int*), so the engine built a base
L4GaugeRenderer whose ctor never parsed gauge/l4gauge.cfg -> empty symbol
table -> "undefined label 'bhk1Init'" -> no ports/gauges. Matched the 3-arg
signature so it truly overrides (args ignored; BT is fullscreen). Same bug
class as the BTL4GaugeRenderer(false,NULL,NULL,NULL) ctor fix.
2. Parse hung on undefined primitives (gated dev accommodation): the BT gauge
primitive table (BTL4MethodDescription) is still a stub, so an unknown
primitive -> ReportParsingError -> Fail() -> a MODAL dialog freezing the
parse. Under BT_DEV_GAUGES, skip the unknown primitive's params so labels
register and the base "configure" primitive builds the ports.
3. Gauge widgets AV on NULL data bindings (gated): NumericDisplayScalar's NULL
value_pointer -> GaugeConnectionDirectOf ctor deref (bind NULL->static zero);
RankAndScore::Execute derefs unreconstructed game state (Gauge::GuardedExecute
SEH wrapper -> Disable(True) on first fault).
All guards gated on BT_DEV_GAUGES: default DEV un-regressed (TARGET DESTROYED,
0 crashes) and the pod path is byte-unchanged (strict Fail, no guards).
Remaining (docs/GAUGE_COMPOSITE.md): the real gauge WIDGET reconstruction
(BTL4MethodDescription method table + gauge->game-state data bindings), then
Step 2 (RP<->BT MFD port-name reconcile) + Step 3 (2-window layout).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PLATFORM PROFILE (-platform pod|dev / BT_PLATFORM / run.cmd pod; default dev):
an env-preset selector, NOT a code fork -- it picks which environment the
existing video/gauge/input code reads. dev = single 800x600 window + keyboard
(the working build); pod = RIO input + (on real hardware) the multi-surface
gauges/MFDs, mirroring content/SETENV.BAT (the pod path -- FindBestAdapterIndices
/ SVGA16 -- is left untouched). The multi-surface needs the pod's 2 video cards,
so -platform pod does NOT auto-enable L4GAUGE on a dev box (stays bootable,
single-window). Also fixes a real engine bug: SVGA16::BuildWindows
PostQuitMessage'd on a CreateDevice failure but fell through to a null-device
deref (segfault) -- now breaks (inert on the pod).
GAUGE DEV-COMPOSITE, Milestone A (option B; opt-in BT_DEV_GAUGES, default OFF):
wake the (dormant) gauge renderer so its CPU-rastered pixelBuffer can later be
composited into the dev window, WITHOUT touching the pod SVGA16 output path.
SVGA16 does no per-surface D3D in this mode (a DevGaugeComposite() gate forces
the no-surface path + short-circuits Update/Refresh). Waking the never-exercised
gauge subsystem exposed 4 latent reconstruction bugs, each guarded:
SVGA16::Update / Refresh (empty mSurfaces[]), LBE4ControlsManager::MakeLinkedLamp
(NULL lamp manager), L4GaugeRenderer::NotifyOfNewInterestingEntity (garbage
warehouse chain). It now boots STABLY; the gauge reconstruction is incomplete
(shadowed lamp-manager + warehouse members) -- Milestone B (the composite pass)
will reveal whether the content is real. See docs/GAUGE_COMPOSITE.md.
Verified: default DEV un-regressed (combat DESTROYED, 0 crashes); pod path
untouched; BT_DEV_GAUGES boots stable.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
