--- id: source-completeness title: "Source Completeness — the gating blocker + the decompilation route" status: established source_sections: "PROGRESS_LOG.md §5a; docs/BT_SOURCE_STATUS.md" related_topics: [wintesla-port, reconstruction-method, decomp-reference, open-questions] key_terms: [BTL4OPT, decomp, oracle, WinTesla, BT] open_questions: - "Backfill the missing 1995 engine headers/TUs (vdata.hpp first) by back-dating the WinTesla/BT412 versions — the compile-closure blocker" --- # Source Completeness — the gating blocker The archive has the **engine and all headers, but NOT the BT game's implementation `.cpp`.** This is THE gating fact that shapes the whole project. Full map: `docs/BT_SOURCE_STATUS.md`. ## What's present vs missing (verified across all 3 zips + the WinTesla archive) - **MUNGA engine** (shared): ~183 `.cpp` — largely complete (renderer, MOVER/JMOVER, math, scene), BUT NOT gapless: a 2026-07-19 header-vs-impl audit + compile bootstrap found missing engine `.cpp` (PLAYER, NETWORK, GRAPH2D, OBJSTRM, SCNROLE, EXPLODE, L4SPLR, L4APP + ~15 smaller) AND at least one missing engine HEADER (`vdata.hpp`, included by plug.hpp — exists only as the drifted WinTesla/BT412 `VDATA.h`). `CODE/BT/MUNGA` is a PARTIAL BT-drift overlay (69 files) over the fuller `CODE/RP/MUNGA` (305); a 1995 build needs the merged BT-over-RP view. [T1] - **BT game `.hpp` headers: PARTIAL, not complete** (2026-07-19 correction — the old "all headers present" claim was WRONG): only 17 of ~36 bt-side `.hpp` survive (mech.hpp, mech2-4, heat.hpp, mechmppr.hpp, torso.hpp, gyro.hpp, hud.hpp, myomers.hpp, powersub.hpp, btplayer.hpp are reconstruction-only), and BTL4VID.HPP/BTL4MPPR.HPP are missing on the btl4 side. `BT.HPP`'s `NO_PRECOMPILED_HEADERS` gate skips the missing mech/btplayer/mechsub includes, which is why TUs outside the mech family still compile-close. [T1] - **BT game logic `.cpp`: MISSING** — `mech`, `mech2/3/4`, `mechsub`, `mechmppr`, `mechtech`, `heat`, `sensor`, `gyro`, `torso`, `myomers`, `hud`, `mechweap`, `powersub`, `btplayer`, and the main app `btl4` are all absent. Only ~10 BT `.cpp` survive (BTCNSL/BTMSSN/BTREG/BTSCNRL/BTTEAM/BTTOOL + GAUSS/PPC + BTL4ARND/BTL4MODE). [T1] - **RP can't backfill BT** — RP's game source is gutted the same way; only shared *engine* game-logic survives. [T1] ## Consequence — routes forward **Route A is CLOSED (2026-07-19): per the user, Nick holds NO assets beyond what's already been provided.** The decompile-and-reconstruct route is THE route: (behavioral) the WinTesla-hosted port in `game/reconstructed/`, and (literal) `TeslaRel410/restoration/source410/` — both fed by the decomp + the manifest. [T1] ## The binary is a near-ideal decompile target `btrel410.exe` is a PKZIP self-extractor; inside is **`BTL4OPT.EXE`** (1.24 MB, clean uncompressed PE32). CORRECTION (2026-07-19 [T1]): the DOS-extender is **Borland PowerPack DPMI32**, not Phar Lap TNT — proven by `OPT.MAK` (`-Tpe`, STARTUP `c0x32.obj + dpmi32.lib + cw32.lib`) + the archived `32STUB.EXE`/`32RTM.EXE`/`DPMI32VM.OVL`. It retains **class names, `Class::Method` strings, and source-file paths in asserts** (`d:\tesla_bt\bt\mech.cpp`) — confirming the missing modules ARE in the binary. No COFF symbol table (naming comes from those strings). Extracted to `reference/decomp/`. [T1] ## THE TOOLCHAIN IS IN HAND (2026-07-19) [T1] **Borland C++ 4.52** identified by BYTE-MATCH: the archived fleet `CW32.LIB` (`CODE/RP`) is identical to the 4.52 install's and differs from 4.5's. Installed at `TeslaRel410/BORLAND/BC45/` (the 4.5 CD was disposed per the user). BCC32/TLINK32/TLIB/MAKE all run NATIVELY on Win11 (hello-world compiled+linked+ran). **`CODE/BT/OPT.MAK` is the shipped binary's exact recipe**: `BCC = bcc32` (NOT bcc32i — that was only EXAMPLE.CFG), `-DLBE4;DEBUG_LEVEL=0;DEBUG_STREAM=cout`, `-5 -a4 -ff -k- -V -Jg -x- -RT- -O2 -Ot -Oc -Og -O -Ol -Z -Ob -Oe -Oi -Om -Op -Ov`, PCH via `-H=btopt.csm`; link `-Tpe -ax` + c0x32/dpmi32/cw32. Gaps: TASM32 absent (separate product; only JOYSTICK.ASM needs it); compile bootstrap of original BTL4MODE.CPP parses deep into the real headers but stops at the missing `vdata.hpp` closure (see What's-missing above). Source paths for a build: include order = `CODE/BT/*` OVER `CODE/RP/*` (BT drift wins), roots per `BTL4.PTH` / `EXAMPLE.CFG` (all under the archive: LIBDPL/SOS/NETNUB in `CODE/BT/MUNGA_L4/`). [T1] ## The mountain, measured (2026-07-18) — the per-TU manifest `tools/manifest410.py` → **`reference/BT410_SOURCE_MANIFEST.md`**: every image function attributed to its original TU via the makefile link order (BT.MAK/BTL4.MAK — census + the tlink32 lib order), Fail() path+line anchors, CLASSMAP seeds, and a monotone-in-link-order DP over data-section cluster affinity. **The missing BT game code = 917 functions / ~230 KB** (bt.lib 482 fn/153 KB + btl4.lib 435 fn/76 KB) out of 6267 image functions; DP extents match every independently-known TU boundary. [T1 region totals / T3 unanchored boundaries] Corrections/leads: the **`.TCP` files are TEST-CLASS files** (debug self-test bodies with SOME shipped code — minor partial survivors for heat, gnrator, emitter, projweap, ammobin, projtile, missile); toolchain = **Borland BCC/tlib/tlink32** (from the makefiles; `TeslaRel410/BORLAND/` placeholder is EMPTY — no byte-match compiler archived yet, though `CODE/RP` keeps C0X32.OBJ/CW32.LIB/32STUB.EXE); `TeslaRel410/sda4` + `CODE/` swept — NO BT `.cpp` there; PVCS conventions (sda4 install) ⇒ **`.CPV`/`.HPV`/`.TCV` are the grep targets** for future drive dumps. Full log: `phases/phase-03-bt410-source-manifest.md`. ## The literal 4.10 tree — `TeslaRel410/restoration/source410/` (started 2026-07-18) The archival-grade recreation of the missing TUs in 1995 house style against the surviving 1995 engine headers (distinct from OUR WinTesla port). Rules + status in its README; every file carries a `.NOTES.md` evidence sidecar; binary `Fail()` file:line records are treated as LINE CONSTRAINTS. **Pilot complete: `BT_L4/BTL4APP.CPP`** — 12/12 functions, all decomp constants resolved to named 1995 expressions (`ApplicationID` BTL4=1 + an unreleased `BTW4` slot in APPMSG.HPP; `ModeInitial`=0x50421; `AddModeMask` inlines byte-exact), Fail lands on its recorded line 400 exactly. [T1/T3 per the sidecar] ## Why reconstruction is viable (THREE anchors) 1. **All BT headers** = exact class layouts, vtable order, signatures (the answer key). 2. **The working WinTesla engine** to compile against ([[wintesla-port]]). 3. **Red Planet's parallel implementations** as a live reference (`VTV`≈mech, `VTVMPPR`≈mechmppr, `WEAPSYS`≈mech weapons, `RP_L4` app≈BT app). Plus `BTL4OPT.EXE` as a behavioral **oracle** (run under cdb / read the decomp). Result = behavior-equivalent **reconstruction**, NOT original source. [T2] ## Key Relationships - Feeds: [[reconstruction-method]] (the loop that fills the gap), [[decomp-reference]]. - Route A closed 2026-07-19 (Nick holds nothing further); gated now only on reconstruction labor + the engine-closure backfill. - Built atop: [[wintesla-port]].