A full comparison against the 175-record ledger found stale guessed coordinates
in both relative slingshot chains. Circles 56/58/71/73 and lines 57/59/72/74
were displaced by up to 70 pixels. The corrected record-57 geometry then
exposed a second issue: readable C and Rust had collapsed two independent
binary response projections into one.
Transcribe the accumulated initializer coordinates exactly. Use the
swapped-component projection only to sign the current tangential response
`abs(normal_velocity) * response_tangent`, and use the standard dot product for
`spin_delta * 0.02 * response_tangent`. Correct the readable C evidence and
Rust together, extend live tracing with the six-byte spin field, and seal the
stopped-Wine transition from `(75530,354765)/(3000,15)` to
`(77369,356597)/(1839,1832)` with Real48 spin `-5.28` in both harnesses.
Test Plan:
- stopped Wine 11.15 record-57 injection/trace -- matched position, velocity, and spin
- `cargo test --workspace --all-targets --all-features` -- 123 passed
- `cargo clippy --workspace --all-targets --all-features -- -D warnings` -- passed
- `bash original/tools/test_reconstructed_c.sh` -- passed
- `python3 original/tools/audit_reconstruction.py --require-complete` -- passed with zero incomplete or unclassified units
- `python3 -m py_compile original/tools/trace_original_state.py original/tools/inject_original_state.py` -- passed
- `rumdl check --flavor commonmark RECONSTRUCTION.md CHANGELOG.md` -- passed
- `git diff --cached --check` -- passed
Replace the partial mechanics transcriptions with a separate, readable C11
reconstruction of the complete Win16 image while preserving the original raw
Ghidra export as immutable evidence. Cover all ordinary and overlapping entry
points, Borland runtime behavior, Win16 imports, segmented data, callbacks,
resources, indirect control flow, physics, rendering, persistence, and
startup/shutdown lifecycles.
Add deterministic extraction and audit tooling plus address-linked ledgers for
functions, imports, DGROUP ranges and objects, relocations, resources, and
callbacks. The final gate records zero raw, partial, restored, unknown,
blocked, or unclassified required units. Keep the semantic-fidelity boundary
explicit: the portable C is not claimed to reproduce a byte-identical Borland
NE build.
Add strict focused harnesses for every reconstructed C unit, exact resource
round-trip checks, and a 16-bit Borland Real48 reference probe. No Rust source
or Cargo metadata is changed in this phase.
Test Plan:
- `bash original/tools/test_reconstructed_c.sh` -- passed
- `bash original/tools/probe_real48_reference.sh` -- passed bit-for-bit
- `python3 original/tools/audit_reconstruction.py --require-complete` -- passed
- `git diff --cached --check` -- passed
- `git diff HEAD -- '*.rs' Cargo.toml Cargo.lock` -- empty
Add a reproducible dump of all 175 initialized collision and rule records from
the unmodified traceable Wine process. Preserve geometry, bounds, activation,
all four decoded Borland Real48 response values, flags, contact state, score,
layer mask, and render rectangle in a reviewable TSV.
This extends the static function ledger with the runtime-built data table needed
to replace inferred target, lock, gate, and mechanism logic in Rust.
Test Plan:
- generated `OBJECTS.tsv` contains one header plus 175 records
- repeat dump compared byte-for-byte equal with `cmp`
- `ruff check` on all original tracing tools -- passed
- `git diff --cached --check` -- passed
Repeated SIGSTOP-based injection can suspend Wine while its Win16 mutex is held
and eventually strand the timer callback. Make the short process-memory write
sequence the default between 30 ms callbacks, while retaining whole-process
stopping as an explicit single-probe diagnostic option.
Test Plan:
- `ruff check original/tools/trace_original_state.py original/tools/inject_original_state.py` -- passed
- live no-stop ball-state injection -- passed
- subsequent original timer updates remained observable
- `git diff --cached --check` -- passed
Expose the original detail-derived integration count, active-ball count,
gravity, maximum speed, and movement flags in live traces. Allow probes to
force one physics substep so a single collision response can be observed
without the default three-step timer batch obscuring intermediate state.
Record the dynamically verified type-2 rail response: resolve velocity in the
registered line's normal/tangent basis, apply the two original Real48
coefficients, then advance from the previous position using that response.
Test Plan:
- `ruff check original/tools/trace_original_state.py original/tools/inject_original_state.py` -- passed
- one-substep probes against both vertical shooter walls -- passed
- ordinary rail responses matched 0.6 normal rebound and 0.1 tangent coupling
- `git diff --cached --check` -- passed
The tracer could only rediscover the game object while all original position
copies agreed, which is normally true at the waiting launcher but not during an
active controlled probe. Accept the injector-reported object base explicitly so
one process can be sampled through an entire sequence of collision tests.
Test Plan:
- `ruff check original/tools/trace_original_state.py original/tools/inject_original_state.py` -- passed
- active live trace with `--object-base 0xadc2b0` -- passed
- `git diff --cached --check` -- passed
Add an atomic live-state injector for the traceable Wine process. It updates
all recovered global and object copies of ball position and velocity while the
process is stopped, clears pending impulses, and re-enables normal collision
processing before resuming the unmodified executable.
This makes individual walls, circles, gates, and mechanisms reachable without
relying on random gameplay. Document the first controlled claw probe, including
its original terminal-18 animation, exact release state, and three-substep
gravity behavior.
Test Plan:
- `ruff check original/tools/trace_original_state.py original/tools/inject_original_state.py` -- passed
- controlled `(289,94)` injection and one-second live claw trace -- passed
- terminal-18 release matched `(325000,92000)` and `(0,1000)` -- passed
- `shellcheck original/tools/run_traceable_original.sh` -- passed
- `git diff --cached --check` -- passed
WineDbg changes the Win16 exception path and cannot capture continuous game
state reliably. Add a Linux-only launcher that opts the normally running Wine
process into same-user tracing without modifying TDKPIN.EXE, plus a sampler
that locates the loaded DGROUP and game object from recovered signatures.
The CSV trace exposes original millipixel positions, predicted positions,
velocities, flipper state, and claw state whenever they change. This provides
an external oracle for reconstructing and differentially testing the fixed-
point mechanics instead of tuning the Rust approximation against screenshots.
Test Plan:
- `shellcheck original/tools/run_traceable_original.sh` -- passed
- `ruff check original/tools/trace_original_state.py` -- passed
- strict shared-library compile with `cc` warnings as errors -- passed
- live Wine smoke trace and four-second charged launch capture -- passed
- `git diff --cached --check` -- passed
ddidderr