BFM-decomp

mirror/BFM-decomp

Fork 0

mirror of https://github.com/Druthulu/BFM-decomp synced 2026-09-27 05:56:00 -04:00

Commit Graph

Author	SHA1	Message	Date
Drew T	070cf375c9	feat(phase-17): T3 — bank 42 sig_unify matches + propagate x134 (fleet 54.48%->55.00%); fix taxonomy classifier - banked 42 gate-proven SIG_FIXABLE_KR matches in ov_SC01_077 (of 48 predicted; 6 honestly reverted by the whole-binary byte-gate, P9) + propagated 13 h_exact-shared x134 via dedup_propagate (one body -> 134 overlays, each byte-gated, fail-closed) - fleet byte-identical-from-source 54.48% -> 55.00% (+0.52%); REAL +1,778; dedup groups 1395 -> 1408; make check-all 136/136 byte-identical (R22); dedup-check 1408 validated/0 failed - wall probe (R14): raw drafts 0/99 + sig_unify fails -> the K&R/sig_unify avenue is exhausted; the residual conflict-class is ARITY_WALL (the §14e param-count dead-end) or structural (T4/T5) - FIX a T1 classifier bug surfaced by verifying the wall vs the bytes: tu_compile_error captured the baseline's pre-existing byte-NEUTRAL warnings ("makes integer from pointer", cc1 rc=0) and mislabeled 99 fns LOOSE_TYPING_WALL; now keeps only non-warning errors -> there is NO loose-typing wall, the real residual wall is ARITY_WALL=74. docs/wall-taxonomy.md amended with the correction. - 134 overlay .c files carry the propagated DEFINE_func instantiations; a one-time full-fleet make extract was required for the per-overlay propagation gate + check-all	2026-06-19 12:13:38 -06:00
Drew T	edb0360d30	feat(phase-17): T1 wall taxonomy — census the 957 residuals, size every avenue (skip T2) - tools/wall_taxonomy.py: full-census classifier (m2c --valid-syntax -> match_one -> sig_unify + full-overlay-TU cc1) labels every unmatched ov_SC01_077 stub by blocker class; reproducible (~22s, --jobs 16); read-only (throwaway _wt_*.c, auto-removed) - docs/wall-taxonomy.md: ranked taxonomy + per-lever aggregation + sequencing recommendation (957 residuals, 99.2% carry an actionable label) - findings (byte-gate-validated, R14: SIG_FIXABLE_KR = exactly the 4/4 gate-passing set): - T4+T5 rich context (struct/jump-table/fn-ptr-table types) = 67% reach, incl. all 30 jtbl giants - T6 permuter = 146 candidates (101 <=4-mismatch near-certain) - T3 sig_unify = 48 gate-proven immediate wins - loose-typing WALL = 99 (10%); non-faithful defer = 16 - T2 DATA-symbol byte-array header = 0 functions -> SKIP (the cookbook §14c-c lever is empty) - recommended resequence: T3 -> T4+T5 (built together) -> T6 ; drop T2 (pending Drew's P5d nod) - struct-core-pivot.md: Phase-17 quantification note; CURRENT_PHASE.md: T1 done + resequence - all binaries byte-identical (T1 read-only; ov_SC01_077 d19c9580 clean rebuild)	2026-06-19 11:31:41 -06:00

Author

SHA1

Message

Date

Drew T

070cf375c9

feat(phase-17): T3 — bank 42 sig_unify matches + propagate x134 (fleet 54.48%->55.00%); fix taxonomy classifier

- banked 42 gate-proven SIG_FIXABLE_KR matches in ov_SC01_077 (of 48 predicted; 6 honestly
  reverted by the whole-binary byte-gate, P9) + propagated 13 h_exact-shared x134 via
  dedup_propagate (one body -> 134 overlays, each byte-gated, fail-closed)
- fleet byte-identical-from-source 54.48% -> 55.00% (+0.52%); REAL +1,778; dedup groups
  1395 -> 1408; make check-all 136/136 byte-identical (R22); dedup-check 1408 validated/0 failed
- wall probe (R14): raw drafts 0/99 + sig_unify fails -> the K&R/sig_unify avenue is exhausted;
  the residual conflict-class is ARITY_WALL (the §14e param-count dead-end) or structural (T4/T5)
- FIX a T1 classifier bug surfaced by verifying the wall vs the bytes: tu_compile_error captured
  the baseline's pre-existing byte-NEUTRAL warnings ("makes integer from pointer", cc1 rc=0) and
  mislabeled 99 fns LOOSE_TYPING_WALL; now keeps only non-warning errors -> there is NO loose-typing
  wall, the real residual wall is ARITY_WALL=74. docs/wall-taxonomy.md amended with the correction.
- 134 overlay .c files carry the propagated DEFINE_func instantiations; a one-time full-fleet
  make extract was required for the per-overlay propagation gate + check-all

2026-06-19 12:13:38 -06:00

Drew T

edb0360d30

feat(phase-17): T1 wall taxonomy — census the 957 residuals, size every avenue (skip T2)

- tools/wall_taxonomy.py: full-census classifier (m2c --valid-syntax -> match_one ->
  sig_unify + full-overlay-TU cc1) labels every unmatched ov_SC01_077 stub by blocker
  class; reproducible (~22s, --jobs 16); read-only (throwaway _wt_*.c, auto-removed)
- docs/wall-taxonomy.md: ranked taxonomy + per-lever aggregation + sequencing recommendation
  (957 residuals, 99.2% carry an actionable label)
- findings (byte-gate-validated, R14: SIG_FIXABLE_KR = exactly the 4/4 gate-passing set):
  - T4+T5 rich context (struct/jump-table/fn-ptr-table types) = 67% reach, incl. all 30 jtbl giants
  - T6 permuter = 146 candidates (101 <=4-mismatch near-certain)
  - T3 sig_unify = 48 gate-proven immediate wins
  - loose-typing WALL = 99 (10%); non-faithful defer = 16
  - T2 DATA-symbol byte-array header = 0 functions -> SKIP (the cookbook §14c-c lever is empty)
- recommended resequence: T3 -> T4+T5 (built together) -> T6 ; drop T2 (pending Drew's P5d nod)
- struct-core-pivot.md: Phase-17 quantification note; CURRENT_PHASE.md: T1 done + resequence
- all binaries byte-identical (T1 read-only; ov_SC01_077 d19c9580 clean rebuild)

2026-06-19 11:31:41 -06:00

2 Commits