mirror of
https://github.com/Druthulu/BFM-decomp
synced 2026-10-07 01:18:49 -04:00
feat(phase-26a): A3f — 33 functions banked that the project had written off as compiler walls
The payoff of A3e, byte-verified. These 33 sat in the backlog at closeness==0 -- match_one said
their bodies were BYTE-EXACT -- and the whole-binary gate rejected them, so they were logged as
`near`/`failed`, i.e. AS MATCHING PROBLEMS, and filed as intrinsic compiler residuals.
They were not hard. They were UNREACHABLE. gate_stage passed `--src src/<ov>/<ov>.c`
unconditionally, which restricts the byte-gate to ONE translation unit -- and every one of these
functions has its stub in a SPLIT TU. Look at where they landed:
src/ov_SC01_077/ov_SC01_077_a.c
src/ov_SC01_077/ov_SC01_077_after.c
src/ov_SC01_077/ov_SC01_077_jr_8012ACE0.c
src/ov_SC01_077/ov_SC01_077_jr_8015AE2C.c
src/ov_SC01_077/ov_SC01_077_jr_8016AB6C.c
src/ov_SC01_077/ov_SC01_077_jr_801734BC.c
src/ov_SC01_077/ov_SC01_077_jr_80178D40.c
src/ov_SC01_077/ov_SC01_077_jr_80182268.c
8 files. SEVEN of them are _jr_/_a/_after carves. NOT ONE is the main .c -- the only file the gate
was ever allowed to look at.
gate: 63 drafts -> banked 33, near 30, FAILED 0 (--no-propagate; the gate and the propagation
are different jobs, and letting an experiment tow an unbounded fleet-wide propagation is
what broke the tree an hour ago)
R22 CLEAN-FLEET: make clean + extract-all + check-all -> 136 passed, 0 failed of 136
dedup-check: 1823 validated, 0 failed | C1 coverage 224933/224933
METRICS, ×1, HONESTLY (no propagation yet -- the multiplier is still ahead):
functions byte-identical 284,526 -> 284,559 (+33)
instructions 8,470,381 -> 8,471,912 (+1,531)
fn-count % 82.79% -> 82.80%
instr-weighted % 66.7% -> 66.7% (flat: ×1 banks do not move the headline)
WHAT THIS MEASURES, beyond the 33: the backlog holds 1,588 entries at closeness==0. 1,215 have been
banked since by other paths. 373 ARE STILL OPEN STUBS WITH BYTE-EXACT BODIES. 63 of them were in
ov_SC01_077 and 33 banked -- a 52% rate on functions the ledger calls unrecoverable. The other 310
are spread across the remaining overlays: same class, same fix, not yet attempted.
Cookbook §51g LAW 11 -- a fix is not landed until its caller stops overriding it. And the reason
this hid for 26 phases, which belongs in the posterity doc: A TOOL THAT CANNOT BANK A FUNCTION IS
INDISTINGUISHABLE, IN EVERY LOG THIS PROJECT KEEPS, FROM A FUNCTION THAT CANNOT BE BANKED.
This commit is contained in:
+1376
-2913
File diff suppressed because it is too large
Load Diff
@@ -4,14 +4,14 @@
|
||||
# cross-binary collapsible-byte leverage: docs/duplicates.cross.md.
|
||||
|
||||
# THREE progress metrics (all matter — see the labels):
|
||||
FLEET fn-count byte-ident: 284526 / 343677 = 82.79% (REAL+LINKED+empties; FUNCTION-count, ×134-inflated — one crack counts per overlay)
|
||||
FLEET instr-weighted : 8470381 / 12707182 = 66.7% (shipped .text across resident+134 overlays; the decomp.dev-DISPLAY number)
|
||||
FLEET distinct-code(uniq): 2530389 / 5410077 = 46.8% (51830/84996 unique fns; the DISTINCT-RE number; main EXE not sig'd)
|
||||
FLEET fn-count byte-ident: 284559 / 343677 = 82.80% (REAL+LINKED+empties; FUNCTION-count, ×134-inflated — one crack counts per overlay)
|
||||
FLEET instr-weighted : 8471912 / 12707182 = 66.7% (shipped .text across resident+134 overlays; the decomp.dev-DISPLAY number)
|
||||
FLEET distinct-code(uniq): 2531879 / 5410077 = 46.8% (51862/84996 unique fns; the DISTINCT-RE number; main EXE not sig'd)
|
||||
|
||||
FLEET REAL substantive : 282998 (of which dedup-shared 224889 via 1823 groups / 224933 instances)
|
||||
FLEET REAL substantive : 283031 (of which dedup-shared 224889 via 1823 groups / 224933 instances)
|
||||
FLEET LINKED PsyQ objs : 959
|
||||
FLEET NON_MATCHING : 7 (0 in any default build — G4)
|
||||
FLEET INCLUDE_ASM stubs : 59144
|
||||
FLEET INCLUDE_ASM stubs : 59111
|
||||
FLEET matchable : 343677
|
||||
|
||||
| binary | REAL | shared | LINKED | byte-ident | matchable | byte-ident % |
|
||||
@@ -26,7 +26,7 @@ FLEET matchable : 343677
|
||||
| ov_SC01_008 | 2080 | 1672 | 0 | 2082 | 2426 | 85.8% |
|
||||
| ov_SC01_009 | 2101 | 1673 | 0 | 2102 | 2507 | 83.8% |
|
||||
| ov_SC01_074 | 2085 | 1673 | 0 | 2087 | 2425 | 86.1% |
|
||||
| ov_SC01_077 | 2320 | 1641 | 0 | 2322 | 2585 | 89.8% |
|
||||
| ov_SC01_077 | 2353 | 1641 | 0 | 2355 | 2585 | 91.1% |
|
||||
| ov_SC01_080 | 2117 | 1676 | 0 | 2117 | 2512 | 84.3% |
|
||||
| ov_SC01_084 | 2124 | 1676 | 0 | 2129 | 2579 | 82.6% |
|
||||
| ov_SC02_000 | 2170 | 1711 | 0 | 2170 | 2683 | 80.9% |
|
||||
|
||||
Reference in New Issue
Block a user