mirror/BFM-decomp
1
0
Fork 0
mirror of https://github.com/Druthulu/BFM-decomp synced 2026-09-26 21:36:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

docs(phase-25): session checkpoint — lock B->A->C order for the fresh session (Drew 2026-07-10c)

Browse Source
This commit is contained in:
Drew T
2026-07-10 21:18:56 -06:00
parent 0241772225
commit c84bcb35c8
1 changed files with 4 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
phase-ends/CURRENT_PHASE.md
+4 -1
View File
@@ -37,7 +37,10 @@ idiom curriculum → sweep one exemplar per family (curriculum-ordered), each cr
deferred mechanical tail (M4 8, F-jumptable 4, M-linkwall 3, M3-residue 4, F-perturb 4, draft 2).
- [ ] Close — clean-fleet verify · PhaseEnd synthesis · plain-English recap (R25) · Phase-26 backlog *(NOT yet — phase open)*
## FRESH SESSION — RESUME HERE (**NEXT (from the frontier map `docs/phase25-frontier-map.md`): (a) the 7 GIANTS 200+ ins → Fable5 surgical probe; (b) the `family_sweep --edit-remap` enhancement (SCOPED DOWN — see recovery block) → recover the 4 byte-drift families func_8016DF5C/80136334/8013D9B0/80156044 + wave-3's func_80136824/80133AB0; (c) F-band backlog — M-linkwall func_8016D1D8/80165240, -O0 func_801457A4, permuter func_8012E364/801549F8**; fleet **74.40%**; phase OPEN, do NOT close)
## FRESH SESSION — RESUME HERE (fleet **74.40%**; phase OPEN, do NOT close. **ORDER for the next sessions, Drew 2026-07-10c: B → A → C.**)
> **B (FIRST) — the `family_sweep --edit-remap` enhancement** (scoped, well-understood — see the recovery block + cookbook §42e). Recover the 6 byte-drift families whose fix lives OUTSIDE the remapped body: **func_8016DF5C, func_80136334, func_8013D9B0, func_80156044** (wave-4) + **func_80136824, func_80133AB0** (wave-3) ≈ **~800 fns**. Approach: {per-sibling — remap the exemplar's `//@EDIT` `old||new` symbols via `family_remap.symbol_map` and apply to the sibling split before the gate} + {once — apply any shared-header (engine_core.h) flip globally, like func_8014FE60/8016CF04}. The `//@EDIT`-carrying candidates are `.run/crack3/wave3/*.c` + `.run/crack3/wave4/*.c` (NOT the `wave*raw` stripped copies). Byte-gate + R22 as always.
> **A (SECOND) — the 7 GIANTS 200+ ins → Fable5 surgical probe** (`.run/frontier_crack_targets.json`, nins≥200: 0x80133cd4/80166994/8014d820/801670e4/80135480/80163ec8/8016cbc0). Prompt Drew for the Fable5 model switch (R27; 7/12 access). **Check `family_remap` remap-ability of each giant up front** (a crack only ×134-propagates if remap-clean or --edit-remap-able — §42e forward rule) so the leverage estimate is honest.
> **C (THIRD) — F-band backlog:** M-linkwall func_8016D1D8/80165240 (compile-fixed, rtu-blind link-fail — manual undefined-syms), sig-reconcile func_80164E40, -O0 func_801457A4 (needs -O0 split handling), permuter func_8012E364 c=4 / func_801549F8 c=3.
> **PROPAGATION-RECOVERY (2026-07-10c, Max) — COMMITTED, R22 136/136 GREEN, fleet 74.36% → 74.40% (+133 fns):**
> Investigated task (b). **R14 CORRECTION:** the "family_remap limitation" I first concluded was a MISDIAGNOSIS — the
> "remap-fail" was a mislabeled **`canon_sig_reconcile` def-finder bug** (its regex required a leading `\n`, so a