fix(ui): render armed record 148 item region

Record 148 publishes player item 20 and refreshes word-quad 20 through overlay
A. Its recovered bounds are (9,14)-(25,30). Rust tracked the armed special hole
but never rendered this visible 17x17 table state.

Expose the armed/active special-hole state and composite the exact DAT997 region
until the special respawn consumes record 148's contact. Make the deterministic
effect-seven scenario internally valid by seeding the required contact, so its
framebuffer now covers both the DAT407 effect and item 20.

Test Plan:
- `cargo test --workspace --all-targets --all-features` -- 120 passed
- `cargo clippy --workspace --all-targets --all-features -- -D warnings` -- passed
- `rumdl check --flavor commonmark RECONSTRUCTION.md CHANGELOG.md` -- passed
- `cargo run -- --simulate effect-7 --step 0 --screenshot /tmp/tdkpin-effect7-special.png` -- passed; visually inspected at 640x460
- launcher/effect screenshot comparison -- 136 RGB pixels differ inside only the expected 17x17 item region
- `git diff --cached --check` -- passed

tdkpin-rs/RECONSTRUCTION.md

@@ -19,7 +19,7 @@ implementation.
 
 | Subsystem | Rust status | Evidence and boundary |
 | --- | --- | --- |
-| Artwork | Exact | All 34 custom DIB images, three standard bitmaps, icon, and palette derivatives are preserved in `assets/original/`. The game uses the original 640x460 table, help, ball, wheel, robot, plunger, media, and item frames. DAT400-407 expose the current per-player bonus/collect/release effect at `(55,79)` while preserving the saved 14x14 wheel overlap. Loading presents DAT995 and advances the original DAT994 strip at raw `1000:531e` destination `(202,328)`, height 13, and clipped width `(stage-1)*7` through stages 2-35. |
+| Artwork | Exact | All 34 custom DIB images, three standard bitmaps, icon, and palette derivatives are preserved in `assets/original/`. The game uses the original 640x460 table, help, ball, wheel, robot, plunger, media, and item frames. DAT400-407 expose the current per-player bonus/collect/release effect at `(55,79)` while preserving the saved 14x14 wheel overlap. Loading presents DAT995 and advances the original DAT994 strip at raw `1000:531e` destination `(202,328)`, height 13, and clipped width `(stage-1)*7` through stages 2-35. Record 148's armed player-item-20 state uses its exact overlay-A Word-Quad `(9,14)-(25,30)`. |
 | Audio | Exact samples and recovered dispatch | All 16 mono PCM WAV resources are embedded unchanged. Playback is emitted at the reconstructed call sites, including multi-sound bank completions, nudge-then-tilt ordering, and Tilt suppression of the pre-reset WAVE-2008 drain cue. Like Win16 `SndPlaySound(SND_ASYNC|SND_NODEFAULT)`, each new resource stops the previous one and plays at the host mixer level without an application-side attenuation. WAV 2022 is loaded by the original generic resource loop but has no playback call and is therefore never played by Rust. |
 | Help and languages | Exact | Original resource images 1001-1005 are displayed directly. |
 | Playfield collision layout | Recovered | All 109 active type-2 line objects and 40 static active type-1 circles are transcribed from the original 175-object registration table. The registration routine converts its sideways inputs with `screen = (y, x - 20)` and accumulates explicitly relative objects. Every static record retains its `+0x49` layer mask; the scanner selects layer 1 below the old Y value 250,000 and layer 2 at or above it, while mask 3 records remain shared. Before detection, the raw `1000:9b69` predicted position must lie in the record bounds derived with the registered five-pixel ball margin. Type-2 records retain every recovered Real48 normal/tangent response pair and registered one-sided orientation. Type-1 records retain their swept-circle radius, radial rebound, tangent coupling, and bumper kick. Detection retains unresolved normal/material candidates so later scan-time motion changes can participate in the selected response. Raw stack slot `SS:...d8a2` makes the first detected record ID win; later nearer candidates are stored but never applied, and both C/Rust have two-candidate regression coverage for this quirk. Each flipper uses its exact two line records plus moving tip circle in both positions. Moving-flipper contact ports `1000:7ed9` rather than fitting live samples: delta-specific pivots/edges, integer cross gates, radial/penetration calculations, response-record gain, and position/velocity publication are tested against all four C harness directions and the raised release geometry. Object 174 is overwritten with the live first ball and Rust handles its ball-to-ball role directly. |