0d6ed40db9784b792ee96ffee0a876b0f342fdc5
9
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Joe DiPrimaandClaude Opus 5
|
4ccc2a7eec |
control mode: Basic re-centre used the STICKY held-button cell -- and the elevation-limit swap was never ported
Sauron: "toggled through advanced controls from standard to advanced and back
to standard -- lost torso control."
The cycle is 0 Basic -> 1 Standard -> 2 Veteran -> WRAPS TO BASIC, so getting
from "advanced" back to Standard PASSES THROUGH BASIC, whose arm re-centres the
torso. @004afbe0 is a complete spec and the port got three things wrong:
iVar1 = mech+0x438 (TORSO) iVar2 = mech+0x5b4 (HUD)
if (mode == 0) { // BASIC
*(iVar1 + 0x1f0) = 0; // analogTwistAxis
*(iVar1 + 0x274) = 1; // recenterActive
*(iVar1 + 0x220) = *(iVar1 + 0x228); // vertLimitTop
*(iVar1 + 0x224) = *(iVar1 + 0x22c); // vertLimitBottom
*(iVar2 + 0x2a0) = 1; // HUD flickerActive
} else if (mode - 1U < 2) { // STANDARD/VETERAN
*(iVar1 + 0x220) = *(iVar1 + 0x230);
*(iVar1 + 0x224) = *(iVar1 + 0x234);
}
1. WRONG CELL. Basic called CommandRecenter() -> centerCommand (@0x208), the
HELD-BUTTON cell: TorsoSimulation re-arms recenterActive from it EVERY frame
it is non-zero, and only the input path clears it -- a mode switch has no
button release to follow. Digital twist commands are processed BEFORE the
centerCommand block, so while it is set they are overridden as fast as they
are applied: the torso stops responding. The binary sets recenterActive
(@0x274) directly -- a ONE-SHOT that self-clears on settle
(`recenterActive = Recenter(dt)`) and is cancelled by any twist input.
2. THE ELEVATION-LIMIT SWAP WAS MISSING ENTIRELY. Two authored pairs exist --
BASIC @0x228/@0x22C (full top, HALF bottom) vs STANDARD/VETERAN @0x230/@0x234
(the full pair) -- and all four members were ctor-written and read by NOTHING.
Basic never restricted downward travel; the assisted modes never restored it.
3. Basic also raises the HUD's flickerActive (@0x2A0) so the horizon re-settles
with the torso it just re-centred. Not ported. (New BTSetHudFlickerActive
bridge in hud.cpp -- mechmppr sees Subsystem*, not HUD.)
Also removed an invented SetAnalogElevationAxis(0); the binary zeroes only 0x1F0.
MEASURED A/B (scratchpad/night13/modecycle.sh, LEGACY=1 for the old path;
BT_LEGACY_MODE_RECENTER=1 is the revert switch):
ctrCmd=1 samples legacy 26 fixed 0
vLim pairs fixed run shows BOTH -- (-0.698..0.349) = -40..20 deg
assisted, and (-0.349..0.349) = -20..20 deg Basic.
Before this commit only the ctor pair ever appeared.
WHAT IS *NOT* PROVEN. I did not reproduce Sauron's PERMANENT loss. In this
bench the legacy latch is periodic, not sticky:
..........LLLL......LLLL......LLLL......LLLL......LLLL......LL
because the desktop key bridge writes centerCommand every frame and zeroes it
when no button is held, so it self-recovers. The torso IS locked while the cell
is set, which is the symptom -- but whether it stays locked depends on the input
path OWNING that cell. On the glass/pad route (Sauron's) nothing may clear it,
which would make it permanent. So: mechanism fixed and binary-grounded, exact
field persistence unverified. Field-verify by cycling modes on a pad build.
Probe: the BT_TORSO_LOG gate line now carries ctrCmd / recen / vLim.
Bench hook: BT_MODECYCLE_EVERY=<n> cycles the mode from the mapper (the pod's
own route is console key 0x13d -- not a RIO button, so BT_BTNTEST cannot press
it, and mech4's BT_MODECYCLE_TEST counter did not advance in a solo run).
NOTE the bench needs BT_KEY_BRIDGE=1: with a PadRIO present the key-bridge
block that consumes the cycle is skipped.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018SgmXGNMXavXiafKXf9MDC
|
||
|
arcattackandClaude Opus 5
|
ef8e449a17 |
#55 steps 0/1/2/4: the 1995 DEATH pass restored, the arg gate made real, and 5 subsystems the respawn sweep never reached
THE HEADLINE: the port's death sweep was a total no-op, and fixing it required
fixing the arg gate in the same commit -- otherwise corpses refill their ammo.
STEP 2 -- the authentic dispatch shape:
* engine/MUNGA/SUBSYSTM.h: Subsystem::DeathShutdown gains the binary's
universal base body { DeathReset(c) } (@004ad10e). It was empty, and NO
port class overrode it, so everything the 1995 game does at death was
skipped entirely.
* mech4.cpp death sweep now passes 0, not 1 (the binary's @0049fe0c arg) --
the wreck shape: alarms/state settle, nothing refills.
* ARG PROPAGATION (mandatory once the sweep runs): AmmoBin::DeathReset is now
arg-gated exactly as @004bd26c -- refill only when arg != 0, ammoAlarm
unconditional. Ignoring the arg was harmless only while the sweep was
dead; with it restored, every corpse would have re-armed. Also forwarded
in PoweredSubsystem / HeatSink / Condenser / HeatableSubsystem / Generator
(each binary body forwards it -- verified addresses in the plan).
STEP 4 -- coverage for 5 classes that had an authentic slot-10 body but no
DeathReset, so the respawn sweep fell through to the empty base:
Torso (this is Gitea #70 -- "loosing torso twist function after a death"),
Gyroscope, Myomers, HUD, Seeker.
STEPS 0/1 -- observability + the David chain:
* Three unconditional matchlog rows: DEAD_NOTIFY (mech + resolved link),
PLAYER_LINK (player <-> vehicle), RESPAWN (mode/alive/zones/subsys/pos).
A respawn was previously INVISIBLE in the matchlog -- night 3's analysis
had to infer them from ammo arithmetic.
* Mech::PlayerLinkMessageHandler + the death dispatch site: when the engine's
one-shot registry lookup misses (no null check, no retry -> the whole
death/respawn cycle silently swallowed), recover the SAME object via the
reverse link the binary's own respawn branch walks (player+0x1FC ==
playerVehicle). Complete-type TU, no raw offsets.
* deathPending cleared in the first-spawn branch (a latch carried in would
permanently kill every later respawn -- the #57 class).
VERIFIED on the 2-pod rig (madcat vs thor, forced kills, 5 death/respawn cycles
per pod): build clean, ZERO new /FORCE unresolved externs from the 5 new
overrides; refill lines appear ONLY immediately before a Mech::Reset and NEVER
between a death and the next respawn (the corpse-refill regression this commit
had to pre-empt); all three probe rows present in both pods' matchlogs; every
DEAD_NOTIFY carried a non-null link.
DELIBERATELY DEFERRED (documented, not forgotten): the mechsub.cpp rename pair
(ResetToInitialState -> GenerateFault, ClearStatus -> the root reset @004ac22c),
deleting HeatableSubsystem::ResetToInitialState, and deleting RespawnRepair.
Those need the HeatableSubsystem vtable-slot-10 pre-flight the plan calls for,
and we have no binary image here to dump the vtable from -- guessing at it
risks the vptr-alias trap. Next session with the decomp shards open.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|
arcattackandClaude Opus 5
|
a3329a64da |
#67 torso replication, part 1: the dead-reckoning clock fix + the bug REPRODUCED on the rig
THE CLOCK FIX (landed, verified live). The replicant torso's extrapolation
window was hardwired to ZERO: both shim accessors returned the same field
(torso.hpp, the "dormant in single-player" note). The engine maintains both
clocks exactly as the binary expects -- WriteUpdateRecord stamps
lastUpdate=lastPerformance on the master, ReadUpdateRecord stamps
lastUpdate=Now() on the replicant (SIMULATE.cpp:276/296, the 1995 "HACK"
comment intact) -- so the fix is one mapping: GetCreationTime (misnamed; now
GetLastUpdateTime) reads lastUpdate. ComputeTargetTwist's
`current - lastUpdate` window is real again. Why legs always replicated
while torsos did not: Mover is ENGINE code reading these fields natively; the
Torso is our reconstruction with the collapsed shim. Rig-verified: the
copy-side log now shows distinct lastUpd/now values.
THE FIELD BUG REPRODUCED (2-pod rig, thor vs thor -- thor because the first
attempt used FOGDAY's Black Hawk, whose torso is authentically FIXED like the
Owens: hEn=0, limits +/-0.01 deg):
[torso-copy] cur=-3.31613 target=-3.31613 atUpd=-3750
The replicant renders at EXACTLY its twist limit (thor: +/-3.31613 rad)
because twistAtUpdate holds garbage (-3750) and the limit clamp slams it to
the stop -- "twisted full right but was not using TT", on demand. An earlier
run poisoned it with 2.1e-44 (= int 15 as float, suspiciously the torso's
subsystem index).
NARROWED: the tx/rx probes added here (BT_TORSO_LOG prints every torso record
both directions, header + raw payload dwords) logged ZERO records in the
poisoned runs -- Torso::Read/WriteUpdateRecord never executed. The garbage
therefore arrives OUTSIDE the torso record path: a raw state write during
mech spawn/consolidation spraying the torso's fields (stream-walker framing
or a consolidation blit). That writer is part 2's hunt; the probes and the
deterministic repro make it a short one.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|
arcattackandClaude Fable 5
|
1271d3bc76 |
Torso elevation (pitch aim) wired: the pod stick's Y axis lives
The mechs could always tilt their aim up/down -- Torso models the full vertical axis (currentElevation, rate, VerticalLimitTop/Bottom, recenter) and EVERY 1995 control mode routes stickPosition.y into Torso::SetAnalogElevationAxis -- but the desktop bridge hard-zeroed stick Y, so the axis was dead on a keyboard rig. The one pod control the remap left unwired. - btinput: JoystickY axis -> elevTarget/elevActive/elevAbsolute - mech4 shim: sElev integrator (same walk/spring model as the twist; X recenters pitch too via gBTElevRecenter) - mechmppr bridge: feeds stickPosition.y every bridged frame (the old unconditional zero removed); both mode branches covered - CONTROLS.MAP (+ numpad profile + compiled default): R/F = aim up/down, pad LeftStickY = elevation (was unused) - torso.hpp: CurrentElevation()/ElevationVelocity() accessors (diag) - [mppr] trace gains stickY (note: the trace reads AFTER the next frame's device push re-zeroes the stick -- input flows regardless) Verified live: R held -> torso elevation climbs at the authored rate and clamps at 0.349066 rad = exactly 20.0 deg (the Blackhawk's VerticalLimitTop); release holds the aim. The eyepoint correctly stays level -- pitch aims the GUNS and reads on the HUD's vertical elevation tape, as in the pod. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> |
||
|
arcattackandClaude Opus 4.8
|
21378ec132 |
Audio Phase 4b (AUDIO_FIDELITY F16): the torso-twist servo whir -- Torso
publishes its real attribute table Recovered the binary Torso attribute table [T1]: dense ids 3..15 from MechSubsystem::NextAttributeID (HeatWatcher/PowerWatcher publish nothing -- binary parity), all 13 rows land on already-reconstructed members: RotationOfTorsoVertical/Horizontal @0x1E4/0x1D8, HorizontalLimitRight/Left @0x1DC/0x1E0, SpeedOfTorsoVertical/Horizontal @0x1EC/0x1E8 (the |rate| the binary abs's -- our derive already did), StickPosition @0x1F0, TorsoUp/Down/ Left/Right/Center @0x1F8..0x208, MotionState @0x20C (statusFlags; the binary writes 2 on the limit-hit frame -- the authored ==2 matcher is the twist-stop clunk, settling the audit's [T4] guess). Torso's AttributeIndex was default-constructed EMPTY -- TorsoTwistInt01/ Ext01/Stop01 were unreachable. Now the authored chain (pitch -200..+200 cents over twist speed 0.5..0.9, start/stop gate at 0.25, stop clunk on MotionState==2) drives them unchanged. Regression (25s): stable; all three bind real; **attrnull count = 0 -- every authored audio attribute in the game now binds a real member.** Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
arcattackandClaude Fable 5
|
02cdfd6576 |
Torso: the TWIST goes LIVE -- electrical watchdog chain, centered crosshair, coherent controls (task #57/#58)
The MadCat torso twists, the view turns with it, and targeting follows. Three reconstruction fronts closed: THE ELECTRICAL WATCHDOG CHAIN (why the torso never powered up): - PowerWatcher::UpdateWatch reconstructed (@004b181c, the REAL registered Performance -- PTR @0050f5fc; Ghidra missed the fn start): the watchdog MIRRORS the watched subsystem's electrical level (+0x278), brownout downgrade when gen output <= minVoltage% x rated. @004b1804 relabeled ResetToInitialState (slot 10) -- the old "Simulation" tag was wrong. - The factory watcher-CONNECT pass reconstructed (vtable slot +0x38, @004aee2c/@004b1a40 byte-identical, recovered from raw exe bytes): watchedLink.Add(roster[watchedSubsystem]) on the master node. Was the SubProxy::Start() no-op -- every watchdog sat at 0 forever. - MinVoltageScale = 0.01 (a 10-byte x87 literal @0x4b1924; was 1.0f = permanent brownout) and PowerWatcher's Derivation chains its REAL base HeatWatcher (the HeatableSubsystem stand-in broke IsDerivedFrom for the whole Torso/Searchlight/ThermalSight family). - KB correction swept: derivation tag 0x50e604 = HEATWATCHER (not "HeatSink"); the btl4gaug heat-widget gate now tests it via the BTIsHeatWatcher bridge. THE CROSSHAIR (task #58 forensics, 6-agent workflow + live probes): - The VIEW is TORSO-MOUNTED: jointtorso -> jointeye -> siteeyepoint in every twist-capable .SKL; the camera + canopy ride the same hinge subtree through HingeRenderable's live matrix-stack compose -- ALREADY WORKING in the port. The crosshair stays screen-centered (center IS the boresight); the twist reads on the tape carets/compass/radar. - The real bug was the port's gBTAimX = tan(twist) slew (the falsified "body-mounted view" model): the camera already carried the twist, so the crosshair counter-slid to hull-forward and the fire ray with it. Deleted; the pick ray inherits the twist from the yawing eye basis. - Two instrumentation traps documented (chase-eye-as-default-camera, BT_FORCE_TORSO clobbering real joints -> the hook now only fills unresolved ones); an over-correcting explicit eye compose was added on those false readings and retired the same day. CONTROLS + REPLICATION: - Q/E spring-center on release (the axis is a twist-RATE demand; the old hold-deflection model drifted forever); X also zeroes the axis and pulses the authentic torso Recenter (@004b6918). M cycles control mode via the real CycleControlMode body. - Torso update-record DIRECTION fixed: engine truth is Write=serialize / Read=apply; @004b6a78 is the READ (was mislabeled Write) and the missing WRITE @004b6a1c recovered from raw disasm (recordLength 0x1C, twist/vel/rate at +0x10/14/18) -- kills the replicant's 0xCDCDCDCD -140-degree ghost twist. - Marching-ghost desync: 4 Standing-case guards zero stale reverse cycleSpeed (negative cadence passed the <= ZeroSpeed stop gate). - Kill credit rerouted to the OBSERVED killer (lastInflictingID -> killer's player link) -- kills count, target K/D populates. KB: subsystems.md (watcher chain), multiplayer.md (record direction), combat-damage.md + gauges-hud.md + cockpit-view.md (torso-mounted view re-correction), decomp-reference.md (new addresses + tag fix), open-questions.md (dead capability-roster loops 2-4, snapshot CD read). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> |
||
|
arcattackandClaude Fable 5
|
cb85517ede |
Gyro: byte-exact re-enable -- ctor map, integrators, jointeye writers (task #56)
The Gyroscope subsystem is LIVE again (was a stub since the NaN revert).
Reconstructed byte-exact from @004b3778 / @004b2ec0 / @004b30ec / @004b33e0 /
@004b34ec + the byte-verified dispatch in the unexported Mech performance
FUN_004a9b5c (calls @0x4aaf74/0x4aaf83):
- ctor field map corrected: springConstant@0x1E8 / dampingConstant@0x1F4 were
mislabelled (the old 'eyeOrientation = r->springConstant' poison line), the
0x254-0x2B3 block (externalPitchPtr, work matrix, placement) was missing so
everything after was mis-offset, and clamps/accumulators were 0xCD fill --
every field now initialised per the binary; 33 offsets static_assert-locked,
sizeof == 0x3D0 exact
- integrators: state-minus-target displacements, componentwise damping
OVERWRITE (FUN_004086d0 is a component multiply, not a cross), position step
without dt, clamp to [posSpring, negSpring]; IntegrateBody's X/Z-crossed
force + integration terms
- writers: WriteMechJoint drives ONE node ('jointeye', type 5): TRANSLATION =
eyePosition spring, ROTATION = bodyOrientation spring; WriteEyeJoint is a
multiplicative sway attenuator on 'jointlocal' (post-anim only)
- dispatch moved out of GyroscopeSimulation into the Mech performance tail
(GyroFrameJointWrite bridge, mech4.cpp) with the binary gates; gyro<->torso
pitch link wired (gyro+0x258 = &torso currentTwist) via complete-type-TU
bridges, retiring the SubProxy::linkTarget landmine
- Mech ctor: deathAnimationLatched/legResetLatch were never initialised --
0xCDCDCDCD gated the writers off silently
Runtime-verified: joints resolve (both BallTranslation), Performance installs,
WriteMechJoint finite from frame 0 (no NaN), Madcat cockpit un-regressed,
combat targeting healthy. Empirical: spring targets are SYMMETRIC -> eye
equilibrium (0,0,0), clamps +/-0.1-0.15u -- the gyro is the hit-BOUNCE
mechanism, not a steady eye offset (hypothesis disproven, recorded in KB).
Pending (task #56 tail): the damage->gyro fan-out FUN_004b2980 (unexported
gap) so hits actually kick the springs; the mech+0x3F0 overspeed sway model;
the torso-pitch EyepointRotation writer (FUN_004b66b4) + glance-look states.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
||
|
arcattackandClaude Opus 4.8
|
d07ac7dd49 |
STEP 6 COMPLETE: cylinder hit-location LIVE — unaimed hits resolve to zones
The Mech per-impact hit-location resolver (the cylinder damage table) is now functional, wired, and runtime-verified [T2]. Unaimed (zone==-1) hits — the collision-damage path — now resolve an impact point to a damage zone via the authentic height x angle grid + weighted dice roll, instead of dropping. dmgtable.cpp/.hpp was a non-functional skeleton on no-op ReconTable/stream shims; backed it with real std::vector storage and fixed 5 latent runtime bugs: - ReadEntries now consumes the leading cell name-string ([i32 len][len+1]) - PieSlice ctor reads rotateWithTorso into the correct member - SelectSlice direct-indexes (was int lookup on a float-keyed table) - ResolveHit returns the zone (chains SelectSlice -> SelectZone) - real MemoryStream::ReadBytes (was a variadic no-op) mech.cpp ctor: replaced the empty-name StandingAnimation stub with the real load — FindResourceDescription(dzRes->resourceName, type 0x1d) -> stream -> new DamageLookupTable, cached at mech[0x111]; ~Mech deletes it. Mech::TakeDamageMessageHandler override registered (MESSAGE_ENTRY overlays Entity's by ID): on invalidDamageZone, resolve via the table then base-route; aimed reticle hits pass through unchanged. Three named accessors (no databinding-trap raw reads): WorldToLocal (localToWorld.MultiplyByInverse), CylinderReferenceHeight (standingTemplateMaxY == collisionTemplate->maxY == binary mech+0x2ec[+0xc]), TorsoHeading via a BTGetTorsoTwist bridge in torso.cpp (Torso::CurrentTwist == torso+0x1d8; torso.hpp cannot be included into mech.cpp — subsystem-stub collision). Stream format + geometry + roll + handler were all byte-verified against the shipped BTL4.RES type-29 resources (18 tables, exact consumption) and the disassembly (FUN_0049eb54/e678/de14, glue 0x49ed0c, handler @0x4a037a). Runtime: boots clean, "[cyl] table 'bhk1' layers=7" (exact byte-verified layer count, found by name), mech spawns + walks, no asserts/AV/0xCDCDCDCD. Env gate BT_CYL_LOG=1. Unblocks collision-damage application. KB updated (combat-damage.md STEP 6 COMPLETE, open-questions.md marked done). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> |
||
|
arcattackandClaude Opus 4.8
|
7b7d465e5e |
Initial commit: bt411 -- standalone Windows BattleTech (Tesla 4.10 port)
Clean, self-contained extraction of the BattleTech-specific work from the
reverse-engineering workspace -- engine + game + content + build, with nothing
from Red Planet or the raw archive dumps. Builds green (Win32) and runs the
single-player drive->animate->target->fire->damage->destroy loop out of the box.
Layout:
engine/ MUNGA + MUNGA_L4 shared 2007 engine, carrying our BT render/loader
work (bgfload/L4D3D/L4VIDEO: BSL bit-slice decode, LOD/ground/shadow
models) + image codec; the minimal rp/ headers the audio HAL needs
game/ reconstructed BT logic + surviving-original BT source + fwd shims
+ WinMain launcher
content/ full runtime tree (BTL4.RES, VIDEO/, GAUGE/, AUDIO/, eggs, BTDPL.INI)
docs/ format specs + reconstruction ledgers
reference/ raw Ghidra pseudocode (recon source-of-truth) + decomp exporter
tools/ MP console emulator + map/resource scanners
One top-level CMake builds munga_engine lib + bt410_l4 game lib + btl4.exe.
All paths relativized (186 fwd shims + ~437 CMake abs paths -> repo-relative);
DXSDK is the one external, overridable via -DDXSDK. Verified: builds to a
byte-identical 2.27MB exe and runs combat (TARGET DESTROYED, 0 crashes) against
the bundled content.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|