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

Files

T

Drew T adafeb13d6 feat(phase-29): T0.3b autopsy — 44% of the "near-miss backlog" are not near-misses; 315 are plumbing-blocked MATCHes

Recomputed every backlog residual from the bytes (1,699 rows, -j 12, zero agent tokens) through the
validated match_one path, deriving asm-subdir + -O0 from corpus.py. R34 cross-check PASSED (closeness
agreed with masked_diff.structured_diff on all 1,610 built rows, 0 classifier errors); 89 nobuild rows
REPORTED not dropped (R32).

BUCKETS: redraft 707 | structural 528 | integration 315 | permuter 57 | unknown 3.

1. HONESTY CORRECTION: 707 of 1,610 (44%) are class SIZE-MISMATCH — the stored best-draft is a
   PARTIAL, an incomplete attempt logged with a closeness score (the func_80183814 666-of-5,122
   shape). docs/backlog.md has been overstating readiness by ~44%. These route to a FRESH CRACK,
   not to a wall and not to the permuter.
2. ACTIONABLE: 315 entries are match_one MATCH *right now*, blocked only on the reconcile ladder —
   recomputing beat trusting the stored label because the tree moved since they were logged.
   ~108,959 gain-ins; top func_80174CB0 (16,482), func_801463A0 (13,534). §52b still applies: ~half
   of close=0 drafts fail the whole-binary gate, so these are CANDIDATES not banks.
3. The permuter bucket is 57/1,610 = 3.5% (Task-13B measured 7.7% and called targeting the problem).
   Extending the mutation set is CONFIRMED not the big lever — small, real, now bounded.
4. R34 again: 3 of 4 comparable labels DISAGREE with measurement — func_80140D68 / func_8012A328 /
   func_801549F8 recorded "schedule" but measure ADDRESSING -> cse. The grinder was aimed wrong.

CONVERGENCE: T0.2 (8/8 failures PLUMBING, 0 walls) and T0.3b (315 integration) independently point at
the SAME lever — the declaration/integration reconcile ladder, worth the 224,410-ins FREE pool AND
~108,959 backlog gain-ins. Two keys eliminated today; untried: canon_sig_reconcile v3.2 and
reconcile_decls.py (the DATA-symbol analog — one T0.2 failure text was a DATA symbol).

2026-07-26 13:25:29 -06:00

4.3 KiB

Raw Blame History

Autopsy — the permuter's failures, classified

Generated by tools/autopsy.py report from .run/autopsy/residuals.jsonl (built by collect, which recompiles every open backlog draft through tools/match_one.py and classifies the FULL residual with tools/residual_class.py).

A verdict here is a candidate, never a bank: the whole-binary byte-gate remains the sole arbiter (G3/P9).

Corpus: 1699 record(s) · 1610 classified · status {'near': 1295, 'nobuild': 89, 'match': 315}

Buckets — what tool each failure actually wants

bucket	fns	reach-weighted	meaning
redraft	707	1638	the stored draft is NOT this function — un-attempted work misfiled as a near-miss; route to a fresh crack, not to a wall
structural	528	5067	local mutation CANNOT introduce it — wants a C-level idiom, not CPU
integration	315	2454	byte-correct standalone; blocked on plumbing (§58/§59 reconcile ladder)
permuter	57	855	local mutation CAN reach it — permuter fuel (a plateau here is a missing-transform candidate)
unknown	3	140	the residue the LLM tier is for

Classes

class	fns	reach-weighted	bucket	permuter profile
SIZE-MISMATCH	707	1638	redraft	—
LENGTH-DRIFT	353	4223	structural	—
MATCH	315	2454	integration	—
OPCODE-MIXED	79	478	structural	—
WIDTH	46	449	structural	—
IMM-OFFSET	38	38	structural	—
BRANCH-POLARITY	21	21	structural	—
ADDRESSING	19	418	permuter	cse
DELAY-SLOT	18	284	permuter	schedule
REGALLOC-PERM	7	7	permuter	regalloc
IMM-VALUE	4	4	permuter	cse
MIXED	3	140	unknown	—

Top clusters (the batch-autopsy work items)

One representative per signature diagnoses the whole cluster — that is what makes the LLM pass a batch autopsy rather than N per-function reads.

#	signature	fns	reach-weighted	example
1	`SIZE-MISMATCH/short`	614	1545	func_8017FF90 (close=4, 2 ins)
2	`MATCH`	315	2454	func_80168B70 (close=0, 27 ins)
3	`LENGTH-DRIFT/-1?`	112	1047	func_8018312C (close=5, 7 ins)
4	`SIZE-MISMATCH/long`	93	93	func_8017E06C (close=6, 9 ins)
5	`LENGTH-DRIFT/-2?`	70	1272	func_80190480 (close=3, 4 ins)
6	`OPCODE-MIXED`	67	466	func_80161D20 (close=1, 14 ins)
7	`LENGTH-DRIFT/1?`	62	461	func_80187074 (close=4, 7 ins)
8	`LENGTH-DRIFT/2?`	50	316	func_80181B8C (close=5, 11 ins)
9	`LENGTH-DRIFT/-7`	32	32	func_80184590 (close=9, 2 ins)
10	`IMM-OFFSET/1`	17	17	func_801827C4 (close=1, 13 ins)
11	`DELAY-SLOT/1`	10	143	func_8016BA68 (close=1, 29 ins)
12	`BRANCH-POLARITY/bnez!=beqz`	10	10	func_80185E38 (close=1, 12 ins)
13	`WIDTH/lhu!=lh`	9	9	func_80142A10 (close=1, 28 ins)
14	`BRANCH-POLARITY/beqz!=bnez`	8	8	func_8017FDB0 (close=1, 12 ins)
15	`OPCODE-MIXED/addressing,width`	6	6	func_8017F114 (close=5, 75 ins)
16	`DELAY-SLOT/2`	6	139	func_80158638 (close=2, 87 ins)
17	`ADDRESSING/lui!=addiu`	6	139	func_8012A328 (close=3, 60 ins)
18	`WIDTH/sh!=lw`	5	5	func_8018AFDC (close=3, 8 ins)
19	`WIDTH/lw!=lhu`	4	270	func_8015FBE0 (close=1, 58 ins)
20	`LENGTH-DRIFT/+1`	4	4	func_8017DCE4 (close=3, 8 ins)

Second oracle — deterministic verdict vs the recorded label

The recorded klass is what a wave agent typed months ago; the verdict is derived from today's bytes. Neither is authoritative — the disagreement is the finding, and it is exactly the mis-classified bucket (a wrongly-labelled residual sends the directed search at the wrong levers, so it plateaus and then reads as a wall).

labelled records comparable: 4 · agree 1 · disagree 3

fn	recorded label	label routes to	measured class	measured routes to
func_80140D68	`schedule`	schedule	ADDRESSING	cse
func_8012A328	`schedule`	schedule	ADDRESSING	cse
func_801549F8	`schedule`	schedule	ADDRESSING	cse

4.3 KiB Raw Blame History