mirror of
https://github.com/Druthulu/BFM-decomp
synced 2026-09-26 13:33:34 -04:00
a8ae190fa3
- src/shared/clearTbl40.h: CLEAR_TBL40 macro = the matched byte-clear loop body, authored ONCE; instantiated at both func_80037004 and func_80037334 in src/800.c (one source -> two vrams) - matched first try; clean rebuild main -> 143dbb89 BYTE-IDENTICAL WITH the shared C AND WITHOUT it (INCLUDE_ASM stub fallback) -> dual invariant proven (R22 clean rebuilds, both states) - config/dedup.us.yaml: I0_clearTbl40 group registered (h_exact a0744d60…); dedup_integrate --check validates it; negative tests (corrupt hash / wrong vram) fail-closed (exit 1) - tools/progress.py: count dedup-shared members as REAL via the registry (the macro form isn't a parseable function def); REAL 52 -> 54, byte-identical 50.24% -> 50.33%; honest measurement (P9) - tools/dedup_integrate.py: display vram in hex in diagnostics - the machinery half of the Phase-11 milestone is proven on the byte-verified EXE