mirror of
https://github.com/Druthulu/BFM-decomp
synced 2026-10-08 17:51:46 -04:00
fix(phase-26a): A3 — the endgame plan was 2.8x too big; the matched set is now DERIVED
docs/family-manifest.md is the document the whole Phase-25/26 structural-family endgame was planned
from. Its matched-set oracle asked ov_SC01_077 ALONE: matched := {h_exact of that one overlay's
non-stub fns} | dedup hashes. So a function ABSENT from that overlay — or stubbed there but matched
in the other 133 — came out "unmatched" and was ranked as live work.
advertised real (derived)
multi-member families 2,758 -> 1,495
"hidden leverage" 11.0 MB -> 3.9 MB
matched-free lever 235/5.9 MB -> 57/1.0 MB
7.1 MB of the advertised leverage was DEAD WORK. And because `instances` counted every overlay
carrying a function — including the ones where it was already banked — the byte-weight RANKING (the
file's entire purpose: "draft these first") was sorted mostly on already-finished code, with the
real targets buried underneath. The A2 audit predicted "true frontier: 1,475 families / 3.9 MB";
derived independently here it is 1,495 / 3.9 MB.
R33: the invariant answers this with no oracle at all —
an h_exact class is WORK iff at least ONE of its instances is still an INCLUDE_ASM stub.
That also makes the dedup-hash union redundant (a dedup-shared member is by definition not a stub),
so the `hash:` regex over config/dedup.us.yaml is DELETED. `instances` now counts only the members
still to bank, so the leverage is the real x-N.
family_hseq: stub scan -> corpus (+100 curated-name stubs the func_-only regex could not see; they
had made 3 still-stubbed functions look like MATCHED exemplars, which every sweep then re-nominates,
produces nothing from, and books as a silent skip). Its hardcoded "expect ~663/~186/~1.85M" self-check
was a stale 2026-07-11 snapshot — 38 banking commits have landed since — and is now labelled a
point-in-time reference, not an invariant. (Verified my change can only GROW the frontier: +100 stubs.)
census_conflict_callees: scoped to src/<ov>/<ov>.c alone, so it saw 13 of 264 stubs and reported
"wave scope: 2 still-stub" when the truth is 57 — every downstream percentage computed against a
denominator 96% too small. Now 0/57 (the audit's exact figure). Its 0-conflict answer was right BY
LUCK; it is now right for a reason. MARKED FOR DELETION (R33): it re-derives from C text what
reconcile_tu.py answers from the build, and its parse holes fail in the UNSAFE direction (an unknown
callee is silently bucketed "conflict-free"). Delete once reconcile_tu is wired into its only consumer.
R14 near-miss, recorded: my first census patch handed collect_stubs() a set of NAMES where it wanted
ADDRESSES, so the membership test was always false and it printed 0/0. Caught only because 0
contradicted the audit's expected 57. A scanner that returns 0 is indistinguishable from a scanner
that found nothing — which is the entire thesis of this audit, and it very nearly bit me while
fixing it.
This commit is contained in:
Drew T
+53
-53
@@ -2,64 +2,64 @@
|
||||
|
||||
> Generated by `tools/family_hseq.py` from the 134 overlay sigs + per-overlay src stubs. Ranked by TEMPLATABLE byte-weight (PURE+IMM members × nins × 4). The byte-gate is the arbiter.
|
||||
|
||||
**Fleet (overlays):** 82.5% fn / 63.6% instr / 40.5% distinct-code matched. Unmatched: 59,680 instances / 4,626,128 ins (34,021 distinct classes).
|
||||
**Fleet (overlays):** 82.8% fn / 66.5% instr / 46.8% distinct-code matched. Unmatched: 58,621 instances / 4,257,491 ins (33,145 distinct classes).
|
||||
|
||||
**Tail cross-check (Phase-25 close):** 42,139 tail fns / 2,299,092 ins → 601 h_seq families ≥2, **179 substantial (nins≥80) / 1,699,870 ins**.
|
||||
**Tail cross-check (Phase-25 close):** 41,366 tail fns / 1,972,283 ins → 597 h_seq families ≥2, **175 substantial (nins≥80) / 1,372,887 ins**.
|
||||
|
||||
**Full frontier (all unmatched by h_seq):** 2730 target families (≥2 members or a matched sibling) + 3553 singletons (Step-D residue). Substantial: **574 families / 3,015,589 templatable ins**, 29 with a matched sibling (zero-crack). Substantial member classes: 16,918 PURE · 39 IMM · 6 STRUCT-excluded.
|
||||
**Full frontier (all unmatched by h_seq):** 2724 target families (≥2 members or a matched sibling) + 3554 singletons (Step-D residue). Substantial: **568 families / 2,646,704 templatable ins**, 30 with a matched sibling (zero-crack). Substantial member classes: 15,855 PURE · 39 IMM · 6 STRUCT-excluded.
|
||||
|
||||
|
||||
## Top substantial families (by templatable byte-weight)
|
||||
|
||||
| # | nins | members (P/I/S) | #addr/#ov | tag | class | matched | jr | exemplar | templ. ins |
|
||||
|--:|--:|--|--|--|--|--:|:-:|--|--:|
|
||||
| 1 | 890 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80178d40 draft-ov077 | 119,260 |
|
||||
| 2 | 952 | 113 (107/6/0) | 21/113 | scattered | IMM | 0 | Y | 0x8017bebc modal | 107,576 |
|
||||
| 3 | 536 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8017a4ac draft-ov077 | 71,824 |
|
||||
| 4 | 493 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015a3c8 draft-ov077 | 66,062 |
|
||||
| 5 | 490 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013f350 draft-ov077 | 65,660 |
|
||||
| 6 | 438 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x801380e0 draft-ov077 | 58,692 |
|
||||
| 7 | 424 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80131340 draft-ov077 | 56,816 |
|
||||
| 8 | 371 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80176734 draft-ov077 | 49,714 |
|
||||
| 9 | 363 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015444c draft-ov077 | 48,642 |
|
||||
| 10 | 337 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80159c84 draft-ov077 | 45,158 |
|
||||
| 11 | 329 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013c414 draft-ov077 | 44,086 |
|
||||
| 12 | 327 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80176218 draft-ov077 | 43,818 |
|
||||
| 13 | 312 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013faf8 draft-ov077 | 41,808 |
|
||||
| 14 | 304 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8014d820 draft-ov077 | 40,736 |
|
||||
| 15 | 289 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80135eb0 draft-ov077 | 38,726 |
|
||||
| 16 | 279 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x801670e4 draft-ov077 | 37,386 |
|
||||
| 17 | 272 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013b83c draft-ov077 | 36,448 |
|
||||
| 18 | 271 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015b950 draft-ov077 | 36,314 |
|
||||
| 19 | 260 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80140958 draft-ov077 | 34,840 |
|
||||
| 20 | 249 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013cb84 draft-ov077 | 33,366 |
|
||||
| 21 | 240 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013d53c draft-ov077 | 32,160 |
|
||||
| 22 | 231 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80175da8 draft-ov077 | 30,954 |
|
||||
| 23 | 226 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8012956c draft-ov077 | 30,284 |
|
||||
| 24 | 222 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015c32c draft-ov077 | 29,748 |
|
||||
| 25 | 213 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013ffd8 draft-ov077 | 28,542 |
|
||||
| 26 | 209 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8016cbc0 draft-ov077 | 28,006 |
|
||||
| 27 | 204 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015ea3c draft-ov077 | 27,336 |
|
||||
| 28 | 201 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015d738 draft-ov077 | 26,934 |
|
||||
| 29 | 198 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013bd74 draft-ov077 | 26,532 |
|
||||
| 30 | 198 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x801412a8 draft-ov077 | 26,532 |
|
||||
| 31 | 188 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015f448 draft-ov077 | 25,192 |
|
||||
| 32 | 188 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8016ab6c draft-ov077 | 25,192 |
|
||||
| 33 | 188 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80175ab8 draft-ov077 | 25,192 |
|
||||
| 34 | 93 | 268 (268/0/0) | 2/134 | cross-address | PURE | 0 | · | 0x8014ffdc draft-ov077 | 24,924 |
|
||||
| 35 | 183 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8014032c draft-ov077 | 24,522 |
|
||||
| 36 | 181 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80135a4c draft-ov077 | 24,254 |
|
||||
| 37 | 174 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8017ae2c draft-ov077 | 23,316 |
|
||||
| 38 | 173 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80154c24 draft-ov077 | 23,182 |
|
||||
| 39 | 165 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80178004 draft-ov077 | 22,110 |
|
||||
| 40 | 162 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80169bbc draft-ov077 | 21,708 |
|
||||
| 41 | 161 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80160534 draft-ov077 | 21,574 |
|
||||
| 42 | 161 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x8016dc20 matched-ov077 | 21,413 |
|
||||
| 43 | 158 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x801299c8 draft-ov077 | 21,172 |
|
||||
| 44 | 156 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x8016df5c matched-ov077 | 20,748 |
|
||||
| 45 | 154 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013c0f8 draft-ov077 | 20,636 |
|
||||
| 46 | 155 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x801365b8 matched-ov077 | 20,615 |
|
||||
| 47 | 154 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x80144090 matched-ov077 | 20,482 |
|
||||
| 48 | 148 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8016d1d8 draft-ov077 | 19,832 |
|
||||
| 49 | 147 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80177b5c draft-ov077 | 19,698 |
|
||||
| 50 | 141 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8014f4c0 draft-ov077 | 18,894 |
|
||||
| 1 | 952 | 112 (106/6/0) | 21/112 | scattered | IMM | 1 | Y | 0x8017bebc matched | 106,624 |
|
||||
| 2 | 536 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8017a4ac draft-ov077 | 71,824 |
|
||||
| 3 | 490 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013f350 draft-ov077 | 65,660 |
|
||||
| 4 | 424 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80131340 draft-ov077 | 56,816 |
|
||||
| 5 | 371 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80176734 draft-ov077 | 49,714 |
|
||||
| 6 | 337 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80159c84 draft-ov077 | 45,158 |
|
||||
| 7 | 329 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013c414 draft-ov077 | 44,086 |
|
||||
| 8 | 327 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80176218 draft-ov077 | 43,818 |
|
||||
| 9 | 312 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013faf8 draft-ov077 | 41,808 |
|
||||
| 10 | 304 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8014d820 draft-ov077 | 40,736 |
|
||||
| 11 | 289 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80135eb0 draft-ov077 | 38,726 |
|
||||
| 12 | 279 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x801670e4 draft-ov077 | 37,386 |
|
||||
| 13 | 272 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013b83c draft-ov077 | 36,448 |
|
||||
| 14 | 271 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015b950 draft-ov077 | 36,314 |
|
||||
| 15 | 260 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80140958 draft-ov077 | 34,840 |
|
||||
| 16 | 249 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013cb84 draft-ov077 | 33,366 |
|
||||
| 17 | 240 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8013d53c draft-ov077 | 32,160 |
|
||||
| 18 | 231 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80175da8 draft-ov077 | 30,954 |
|
||||
| 19 | 226 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8012956c draft-ov077 | 30,284 |
|
||||
| 20 | 222 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8015c32c draft-ov077 | 29,748 |
|
||||
| 21 | 209 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8016cbc0 draft-ov077 | 28,006 |
|
||||
| 22 | 204 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015ea3c draft-ov077 | 27,336 |
|
||||
| 23 | 201 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015d738 draft-ov077 | 26,934 |
|
||||
| 24 | 198 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013bd74 draft-ov077 | 26,532 |
|
||||
| 25 | 198 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x801412a8 draft-ov077 | 26,532 |
|
||||
| 26 | 188 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8015f448 draft-ov077 | 25,192 |
|
||||
| 27 | 188 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80175ab8 draft-ov077 | 25,192 |
|
||||
| 28 | 93 | 268 (268/0/0) | 2/134 | cross-address | PURE | 0 | · | 0x8014ffdc draft-ov077 | 24,924 |
|
||||
| 29 | 183 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8014032c draft-ov077 | 24,522 |
|
||||
| 30 | 181 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80135a4c draft-ov077 | 24,254 |
|
||||
| 31 | 174 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8017ae2c draft-ov077 | 23,316 |
|
||||
| 32 | 173 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80154c24 draft-ov077 | 23,182 |
|
||||
| 33 | 165 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80178004 draft-ov077 | 22,110 |
|
||||
| 34 | 162 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80169bbc draft-ov077 | 21,708 |
|
||||
| 35 | 161 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80160534 draft-ov077 | 21,574 |
|
||||
| 36 | 161 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x8016dc20 matched-ov077 | 21,413 |
|
||||
| 37 | 158 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x801299c8 draft-ov077 | 21,172 |
|
||||
| 38 | 156 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x8016df5c matched-ov077 | 20,748 |
|
||||
| 39 | 154 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x8013c0f8 draft-ov077 | 20,636 |
|
||||
| 40 | 155 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x801365b8 matched-ov077 | 20,615 |
|
||||
| 41 | 154 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x80144090 matched-ov077 | 20,482 |
|
||||
| 42 | 148 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8016d1d8 draft-ov077 | 19,832 |
|
||||
| 43 | 147 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80177b5c draft-ov077 | 19,698 |
|
||||
| 44 | 141 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8014f4c0 draft-ov077 | 18,894 |
|
||||
| 45 | 141 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x8013d9b0 matched-ov077 | 18,753 |
|
||||
| 46 | 139 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x8014ade0 draft-ov077 | 18,626 |
|
||||
| 47 | 136 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x80135260 draft-ov077 | 18,224 |
|
||||
| 48 | 137 | 133 (133/0/0) | 1/133 | per-location | PURE | 1 | · | 0x80133ab0 matched-ov077 | 18,221 |
|
||||
| 49 | 131 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | Y | 0x801555f4 draft-ov077 | 17,554 |
|
||||
| 50 | 129 | 134 (134/0/0) | 1/134 | per-location | PURE | 0 | · | 0x80137dd4 draft-ov077 | 17,286 |
|
||||
|
||||
+62
-62
@@ -2,69 +2,69 @@
|
||||
|
||||
> Generated by `tools/family_manifest.py` from the 134 overlay sigs + ov_SC01_077 matched set + `config/dedup.us.yaml`. One exemplar per h_norm family, drafted from its **ov_SC01_077 member** where present. h_norm families are reloc-only by construction; whether one C body matches all members is confirmed per-family by the byte-gate (T3 magnitude / T7 banks). **Leverage = `inst` (fleet-wide member count), NOT `hexR` (h_exact-reach, ~1-3 for byte-shattered families).**
|
||||
|
||||
**Families:** 44101 · **multi-member (hidden leverage):** 2758 / 11.0 MB · **with an ov_SC01_077 draftable exemplar:** 42.
|
||||
**Families:** 23633 · **multi-member (hidden leverage):** 1495 / 3.9 MB · **with an ov_SC01_077 draftable exemplar:** 17.
|
||||
|
||||
**Endgame levers** (multi-member families × ov_SC01_077 membership): **draftable** 42 / 3.0 MB (unmatched ov077 member → draft the exemplar, T5/T7) · **matched-free** 235 / 5.9 MB (matched ov077 member + unmatched siblings → free `--tier h_norm` propagation, T3) · **absent** 2481 / 2.1 MB (no ov077 member → draft from another overlay).
|
||||
**Endgame levers** (multi-member families × ov_SC01_077 membership): **draftable** 17 / 1.1 MB (unmatched ov077 member → draft the exemplar, T5/T7) · **matched-free** 57 / 1.0 MB (matched ov077 member + unmatched siblings → free `--tier h_norm` propagation, T3) · **absent** 1421 / 1.9 MB (no ov077 member → draft from another overlay).
|
||||
|
||||
| # | rep addr | ov077 exemplar | nins | hexcls | inst | hexR | byteMB | bucket |
|
||||
|--:|----------|----------------|-----:|-------:|-----:|-----:|-------:|--------|
|
||||
| 1 | 0x80133cd4 | 0x80133cd4 | 399 | 127 | 134 | 3 | 0.21 | GIANT |
|
||||
| 2 | 0x80166994 | 0x80166994 | 369 | 127 | 134 | 3 | 0.20 | GIANT |
|
||||
| 3 | 0x8013c414 | 0x8013c414 | 329 | 127 | 134 | 3 | 0.18 | GIANT |
|
||||
| 4 | 0x80150480 | 0x8014fbc0 | 22 | 1891 | 1996 | 3 | 0.18 | small |
|
||||
| 5 | 0x8014d820 | 0x8014d820 | 304 | 126 | 134 | 3 | 0.16 | GIANT |
|
||||
| 6 | 0x8016a290 | — | 284 | 125 | 133 | 3 | 0.15 | GIANT |
|
||||
| 7 | 0x801670e4 | 0x801670e4 | 279 | 126 | 134 | 3 | 0.15 | GIANT |
|
||||
| 8 | 0x80130d48 | — | 266 | 125 | 133 | 3 | 0.14 | GIANT |
|
||||
| 9 | 0x80135480 | 0x80135480 | 258 | 127 | 134 | 3 | 0.14 | GIANT |
|
||||
| 10 | 0x80163ec8 | 0x80163ec8 | 234 | 127 | 134 | 3 | 0.13 | GIANT |
|
||||
| 11 | 0x80134c20 | — | 230 | 126 | 133 | 3 | 0.12 | GIANT |
|
||||
| 12 | 0x8016cbc0 | 0x8016cbc0 | 209 | 127 | 134 | 3 | 0.11 | GIANT |
|
||||
| 13 | 0x80133784 | — | 203 | 126 | 133 | 3 | 0.11 | GIANT |
|
||||
| 14 | 0x8013bd74 | 0x8013bd74 | 198 | 127 | 134 | 3 | 0.11 | GIANT |
|
||||
| 15 | 0x80166690 | 0x80166690 | 193 | 127 | 134 | 3 | 0.10 | GIANT |
|
||||
| 16 | 0x80150170 | 0x8014fe60 | 95 | 2 | 268 | 134 | 0.10 | med |
|
||||
| 17 | 0x8013b274 | — | 189 | 125 | 133 | 3 | 0.10 | GIANT |
|
||||
| 18 | 0x8013dd68 | — | 187 | 126 | 133 | 3 | 0.10 | GIANT |
|
||||
| 19 | 0x80167ae0 | — | 183 | 125 | 133 | 3 | 0.10 | GIANT |
|
||||
| 20 | 0x80161e08 | — | 175 | 126 | 133 | 3 | 0.09 | GIANT |
|
||||
| 21 | 0x80167dbc | — | 173 | 125 | 133 | 3 | 0.09 | GIANT |
|
||||
| 22 | 0x8016c49c | — | 172 | 126 | 133 | 3 | 0.09 | GIANT |
|
||||
| 23 | 0x8016cf04 | 0x8016cf04 | 166 | 127 | 134 | 3 | 0.09 | GIANT |
|
||||
| 24 | 0x80136334 | 0x80136334 | 161 | 126 | 134 | 3 | 0.09 | GIANT |
|
||||
| 25 | 0x801347a0 | — | 162 | 125 | 133 | 3 | 0.09 | GIANT |
|
||||
| 26 | 0x8016dc20 | — | 161 | 126 | 133 | 3 | 0.09 | GIANT |
|
||||
| 27 | 0x80162438 | — | 160 | 126 | 133 | 3 | 0.09 | GIANT |
|
||||
| 28 | 0x8016df5c | 0x8016df5c | 156 | 127 | 134 | 3 | 0.08 | GIANT |
|
||||
| 29 | 0x801365b8 | 0x801365b8 | 155 | 126 | 134 | 3 | 0.08 | GIANT |
|
||||
| 30 | 0x80136824 | — | 156 | 125 | 133 | 3 | 0.08 | GIANT |
|
||||
| 31 | 0x8013c0f8 | 0x8013c0f8 | 154 | 127 | 134 | 3 | 0.08 | GIANT |
|
||||
| 32 | 0x8016d1d8 | 0x8016d1d8 | 148 | 127 | 134 | 3 | 0.08 | large |
|
||||
| 33 | 0x8012c890 | — | 149 | 125 | 133 | 3 | 0.08 | large |
|
||||
| 34 | 0x8013d9b0 | 0x8013d9b0 | 141 | 127 | 134 | 3 | 0.08 | large |
|
||||
| 35 | 0x80133ab0 | — | 137 | 126 | 133 | 3 | 0.07 | large |
|
||||
| 36 | 0x80140f00 | — | 128 | 125 | 133 | 3 | 0.07 | large |
|
||||
| 37 | 0x80145cec | — | 127 | 125 | 133 | 3 | 0.07 | large |
|
||||
| 38 | 0x8017d960 | — | 3338 | 5 | 5 | 1 | 0.07 | GIANT |
|
||||
| 39 | 0x8013eb7c | — | 124 | 125 | 133 | 3 | 0.07 | large |
|
||||
| 40 | 0x80145a2c | — | 62 | 250 | 266 | 3 | 0.07 | med |
|
||||
| 41 | 0x8013ad38 | — | 122 | 125 | 133 | 3 | 0.06 | large |
|
||||
| 42 | 0x8013d178 | — | 110 | 126 | 133 | 3 | 0.06 | large |
|
||||
| 43 | 0x8013221c | 0x80132144 | 27 | 506 | 539 | 3 | 0.06 | small |
|
||||
| 44 | 0x8013ef88 | — | 108 | 125 | 133 | 3 | 0.06 | large |
|
||||
| 45 | 0x80168828 | — | 108 | 125 | 133 | 3 | 0.06 | large |
|
||||
| 46 | 0x80134a74 | — | 107 | 126 | 133 | 3 | 0.06 | large |
|
||||
| 47 | 0x801345f8 | — | 106 | 126 | 133 | 3 | 0.06 | large |
|
||||
| 48 | 0x80141100 | — | 106 | 125 | 133 | 3 | 0.06 | large |
|
||||
| 49 | 0x8017b614 | — | 101 | 126 | 133 | 3 | 0.05 | large |
|
||||
| 50 | 0x8017b490 | 0x8017b490 | 97 | 127 | 134 | 3 | 0.05 | med |
|
||||
| 51 | 0x8013dbe4 | — | 97 | 126 | 133 | 3 | 0.05 | med |
|
||||
| 52 | 0x801596f0 | — | 97 | 125 | 133 | 3 | 0.05 | med |
|
||||
| 53 | 0x8016bfd0 | — | 95 | 125 | 133 | 3 | 0.05 | med |
|
||||
| 54 | 0x8013ee10 | — | 94 | 125 | 133 | 3 | 0.05 | med |
|
||||
| 55 | 0x801418f8 | 0x801418f8 | 90 | 127 | 134 | 3 | 0.05 | med |
|
||||
| 56 | 0x8013d3d4 | — | 90 | 126 | 133 | 3 | 0.05 | med |
|
||||
| 57 | 0x8013e6ac | — | 90 | 125 | 133 | 3 | 0.05 | med |
|
||||
| 58 | 0x80131b14 | 0x80131b14 | 89 | 126 | 134 | 3 | 0.05 | med |
|
||||
| 59 | 0x80135004 | — | 89 | 126 | 133 | 3 | 0.05 | med |
|
||||
| 60 | 0x801397b0 | — | 89 | 125 | 133 | 3 | 0.05 | med |
|
||||
| 1 | 0x8013c414 | 0x8013c414 | 329 | 127 | 134 | 3 | 0.18 | GIANT |
|
||||
| 2 | 0x8014d820 | 0x8014d820 | 304 | 126 | 134 | 3 | 0.16 | GIANT |
|
||||
| 3 | 0x801670e4 | 0x801670e4 | 279 | 126 | 134 | 3 | 0.15 | GIANT |
|
||||
| 4 | 0x8016cbc0 | 0x8016cbc0 | 209 | 127 | 134 | 3 | 0.11 | GIANT |
|
||||
| 5 | 0x8013bd74 | 0x8013bd74 | 198 | 127 | 134 | 3 | 0.11 | GIANT |
|
||||
| 6 | 0x8016dc20 | — | 161 | 126 | 133 | 3 | 0.09 | GIANT |
|
||||
| 7 | 0x8016df5c | — | 156 | 126 | 133 | 3 | 0.08 | GIANT |
|
||||
| 8 | 0x8013c0f8 | 0x8013c0f8 | 154 | 127 | 134 | 3 | 0.08 | GIANT |
|
||||
| 9 | 0x801365b8 | — | 155 | 125 | 133 | 3 | 0.08 | GIANT |
|
||||
| 10 | 0x8016d1d8 | 0x8016d1d8 | 148 | 127 | 134 | 3 | 0.08 | large |
|
||||
| 11 | 0x8013d9b0 | — | 141 | 126 | 133 | 3 | 0.08 | large |
|
||||
| 12 | 0x80133ab0 | — | 137 | 126 | 133 | 3 | 0.07 | large |
|
||||
| 13 | 0x8017d960 | — | 3338 | 5 | 5 | 1 | 0.07 | GIANT |
|
||||
| 14 | 0x8013d178 | — | 110 | 126 | 133 | 3 | 0.06 | large |
|
||||
| 15 | 0x8013ef88 | — | 108 | 125 | 133 | 3 | 0.06 | large |
|
||||
| 16 | 0x80134a74 | — | 107 | 126 | 133 | 3 | 0.06 | large |
|
||||
| 17 | 0x8017b490 | — | 97 | 126 | 133 | 3 | 0.05 | med |
|
||||
| 18 | 0x801418f8 | — | 90 | 126 | 133 | 3 | 0.05 | med |
|
||||
| 19 | 0x80143d28 | — | 80 | 125 | 133 | 3 | 0.04 | med |
|
||||
| 20 | 0x801457a4 | 0x801457a4 | 79 | 126 | 134 | 3 | 0.04 | med |
|
||||
| 21 | 0x8016c188 | — | 79 | 126 | 133 | 3 | 0.04 | med |
|
||||
| 22 | 0x80141a60 | — | 76 | 125 | 133 | 3 | 0.04 | med |
|
||||
| 23 | 0x80156044 | — | 74 | 125 | 133 | 3 | 0.04 | med |
|
||||
| 24 | 0x8012e364 | 0x8012e364 | 67 | 127 | 134 | 3 | 0.04 | med |
|
||||
| 25 | 0x80165240 | 0x80165240 | 63 | 126 | 134 | 3 | 0.03 | med |
|
||||
| 26 | 0x8016d688 | 0x8016d688 | 60 | 127 | 134 | 3 | 0.03 | med |
|
||||
| 27 | 0x801801e8 | — | 195 | 34 | 36 | 2 | 0.03 | GIANT |
|
||||
| 28 | 0x8013d8fc | — | 45 | 126 | 133 | 3 | 0.02 | small |
|
||||
| 29 | 0x8017da90 | — | 227 | 25 | 26 | 2 | 0.02 | GIANT |
|
||||
| 30 | 0x80186b78 | — | 236 | 20 | 20 | 1 | 0.02 | GIANT |
|
||||
| 31 | 0x801549f8 | 0x801549f8 | 31 | 127 | 134 | 3 | 0.02 | small |
|
||||
| 32 | 0x8017cd9c | — | 793 | 5 | 5 | 1 | 0.02 | GIANT |
|
||||
| 33 | 0x80180a64 | — | 766 | 3 | 5 | 2 | 0.02 | GIANT |
|
||||
| 34 | 0x8017d8b8 | 0x8017c294 | 246 | 6 | 15 | 5 | 0.01 | GIANT |
|
||||
| 35 | 0x8017cdd8 | — | 113 | 28 | 32 | 2 | 0.01 | large |
|
||||
| 36 | 0x80180324 | — | 328 | 11 | 11 | 1 | 0.01 | GIANT |
|
||||
| 37 | 0x80187e4c | — | 237 | 15 | 15 | 1 | 0.01 | GIANT |
|
||||
| 38 | 0x8017c9bc | — | 253 | 13 | 14 | 2 | 0.01 | GIANT |
|
||||
| 39 | 0x80164e40 | 0x80164e40 | 25 | 127 | 134 | 3 | 0.01 | small |
|
||||
| 40 | 0x801808e0 | — | 92 | 11 | 36 | 10 | 0.01 | med |
|
||||
| 41 | 0x80180034 | — | 90 | 11 | 36 | 10 | 0.01 | med |
|
||||
| 42 | 0x8017c43c | — | 293 | 10 | 11 | 2 | 0.01 | GIANT |
|
||||
| 43 | 0x8013bc7c | — | 24 | 125 | 133 | 3 | 0.01 | small |
|
||||
| 44 | 0x8017efb0 | — | 223 | 14 | 14 | 1 | 0.01 | GIANT |
|
||||
| 45 | 0x8017f510 | — | 1511 | 2 | 2 | 1 | 0.01 | GIANT |
|
||||
| 46 | 0x8013bcdc | — | 22 | 125 | 133 | 3 | 0.01 | small |
|
||||
| 47 | 0x8017ff08 | — | 263 | 11 | 11 | 1 | 0.01 | GIANT |
|
||||
| 48 | 0x80185150 | — | 185 | 13 | 15 | 2 | 0.01 | GIANT |
|
||||
| 49 | 0x8017f9bc | — | 386 | 6 | 7 | 2 | 0.01 | GIANT |
|
||||
| 50 | 0x80188e6c | — | 168 | 15 | 16 | 2 | 0.01 | GIANT |
|
||||
| 51 | 0x801849ac | — | 177 | 13 | 15 | 2 | 0.01 | GIANT |
|
||||
| 52 | 0x8017e1dc | — | 96 | 25 | 26 | 2 | 0.01 | med |
|
||||
| 53 | 0x8017d1a0 | — | 10 | 237 | 243 | 3 | 0.01 | tiny |
|
||||
| 54 | 0x8013b7f4 | — | 18 | 126 | 133 | 3 | 0.01 | tiny |
|
||||
| 55 | 0x8017cdd8 | — | 17 | 116 | 137 | 3 | 0.01 | tiny |
|
||||
| 56 | 0x80188c28 | — | 145 | 15 | 16 | 2 | 0.01 | large |
|
||||
| 57 | 0x80188440 | — | 154 | 15 | 15 | 1 | 0.01 | GIANT |
|
||||
| 58 | 0x80180474 | — | 164 | 14 | 14 | 1 | 0.01 | GIANT |
|
||||
| 59 | 0x80188850 | — | 147 | 15 | 15 | 1 | 0.01 | large |
|
||||
| 60 | 0x80188200 | — | 144 | 15 | 15 | 1 | 0.01 | large |
|
||||
|
||||
@@ -21,6 +21,9 @@ import argparse, json, os, importlib.util
|
||||
from collections import defaultdict
|
||||
|
||||
REPO = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
|
||||
import sys
|
||||
sys.path.insert(0, os.path.join(REPO, 'tools'))
|
||||
import corpus # the derived corpus oracle (Phase 26-A)
|
||||
|
||||
_spec = importlib.util.spec_from_file_location('ght', os.path.join(REPO, 'tools/gen_harvest_targets.py'))
|
||||
_ght = importlib.util.module_from_spec(_spec)
|
||||
@@ -36,15 +39,32 @@ def main():
|
||||
args = ap.parse_args()
|
||||
|
||||
src = args.source
|
||||
c_path = os.path.join(REPO, f'src/{src}/{src}.c')
|
||||
ec = os.path.join(REPO, 'src/shared/engine_core.h')
|
||||
|
||||
# ⚠ DEPRECATED — SCHEDULED FOR DELETION (Phase 26-A audit, R33).
|
||||
# This tool re-derives, by re-parsing C text, the question "for this TU, which callees are
|
||||
# defined / declared / stubbed / external?" — which is exactly and only what tools/reconcile_tu.py
|
||||
# answers FROM THE BUILD. A parse hole in reconcile_tu makes it CONSERVATIVE; a parse hole here
|
||||
# makes it WRONG IN THE UNSAFE DIRECTION (an unknown callee is silently bucketed "conflict-free").
|
||||
# Delete this file once reconcile_tu is wired into derive_canonical_sigs (its only consumer).
|
||||
#
|
||||
# Until then, at least stop the 96% under-report: the corpus was `src/<ov>/<ov>.c` ALONE, so it
|
||||
# saw 13 of ov_SC01_077's 264 stubs and printed "wave scope: 2 still-stub" when the truth is 57 —
|
||||
# every downstream percentage was computed against a denominator 96% too small. The 0-conflict
|
||||
# answer it gives today is right BY LUCK (those callees happen to be banked); point it at a fresh
|
||||
# overlay and it would see ~14 of ~600 stubs and confidently report zero conflicts into a wave
|
||||
# riddled with them — the exact failure that cost 60%-vs-33% in the Phase-17 calibration.
|
||||
c_paths = [str(p) for p in corpus.src_files(src)]
|
||||
c_path = c_paths[0] if c_paths else os.path.join(REPO, f'src/{src}/{src}.c')
|
||||
|
||||
src_sig = _ght.load_sig(os.path.join(REPO, f'.run/sig.{src}.jsonl')) # addr-int -> {calls,nins,h_exact,reach?}
|
||||
stubs = _ght.collect_stubs(c_path) # current INCLUDE_ASM set
|
||||
stubs = set(corpus.stubs(src)) # ADDRESSES (ints) — ALL TUs, ANY symbol name
|
||||
define_sigs = _ght.collect_define_sigs(ec)
|
||||
inline_sigs = _ght.collect_inline_sigs(c_path)
|
||||
inline_sigs = {}
|
||||
for cp in c_paths:
|
||||
inline_sigs.update(_ght.collect_inline_sigs(cp))
|
||||
defined = {**define_sigs, **inline_sigs}
|
||||
extern_sigs = _ght.collect_extern_sigs([ec, c_path])
|
||||
extern_sigs = _ght.collect_extern_sigs([ec] + c_paths)
|
||||
|
||||
s3 = json.load(open(os.path.join(REPO, args.targets)))
|
||||
tgt_addrs = [int(t['addr'], 16) for t in s3]
|
||||
|
||||
+14
-7
@@ -20,7 +20,7 @@ arbiter (G3/P9): this survey RANKS and CLASSES; it never asserts a match.
|
||||
|
||||
Ground truth = the sigs (`make sig-overlays`) + src stubs. Companion to tools/family_manifest.py (h_norm).
|
||||
"""
|
||||
import json, glob, re, collections, sys
|
||||
import os, sys, json, glob, re, collections, sys
|
||||
sys.path.insert(0, "tools")
|
||||
import family_remap as FR
|
||||
|
||||
@@ -29,13 +29,18 @@ TINY = 16 # nins < this is the coincidental-h_seq-collision band
|
||||
EX_OV = "ov_SC01_077" # the canonical drafting overlay (cached Ghidra-C)
|
||||
|
||||
|
||||
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
||||
import corpus # the derived corpus oracle (Phase 26-A)
|
||||
|
||||
|
||||
def load():
|
||||
"""-> instances[(ov, addr, nins, h_exact, h_norm, h_seq, matched)], and nins/matched maps."""
|
||||
stubs = {}
|
||||
for ovdir in sorted(glob.glob("src/ov_*")):
|
||||
ov = ovdir.split("/")[-1]
|
||||
txt = "".join(open(f).read() for f in glob.glob(f"{ovdir}/*.c"))
|
||||
stubs[ov] = set(int(m, 16) for m in re.findall(r'INCLUDE_ASM\([^)]*,\s*func_([0-9A-Fa-f]+)\)', txt))
|
||||
# DERIVED from tools/corpus.py (Phase 26-A). The old scan globbed every .c correctly but its
|
||||
# regex only matched `func_<hex>` symbols — so the 100 CURATED-name stubs (listCdBuffer) were
|
||||
# invisible, and family_hseq therefore labelled 3 still-stubbed functions as MATCHED exemplars.
|
||||
# A phantom exemplar is re-nominated by every sweep, produces nothing, and books a silent skip.
|
||||
stubs = {ov: set(corpus.stubs(ov)) for ov in
|
||||
(d.split("/")[-1] for d in sorted(glob.glob("src/ov_*")))}
|
||||
inst = []
|
||||
for p in sorted(glob.glob(".run/sig.ov_*.jsonl")):
|
||||
ov = f"ov_{p.split('sig.ov_')[1][:-6]}"
|
||||
@@ -197,7 +202,9 @@ def main():
|
||||
print(f"fleet: {metrics['fn_count_matched_pct']}% fn / {metrics['instr_weighted_matched_pct']}% instr / "
|
||||
f"{metrics['distinct_matched_pct']}% distinct")
|
||||
print(f"tailcheck: {tailcheck['hseq_families_ge2']} families ≥2, {tailcheck['substantial_families']} "
|
||||
f"substantial / {tailcheck['substantial_ins']:,} ins (expect ~663 / ~186 / ~1.85M)")
|
||||
f"substantial / {tailcheck['substantial_ins']:,} ins"
|
||||
f" [Task-2 snapshot 2026-07-11: 663 / 186 / 1.85M — this number SHRINKS as banking "
|
||||
f"proceeds; it is a point-in-time reference, NOT an invariant to match]")
|
||||
print(f"frontier: {len(multi)} target families ({len(subst)} substantial, {len(with_matched)} w/ matched sib) "
|
||||
f"+ {len(singles)} singletons")
|
||||
print(f"-> .run/family_hseq.json + docs/family-hseq.md")
|
||||
|
||||
+38
-18
@@ -15,7 +15,10 @@ many overlays carry it), NOT h_exact-reach (which is ~1-3 for a byte-shattered f
|
||||
Ground truth = the sigs (they sign the ORIGINAL bytes, stable across matching); reproducible via
|
||||
`make sig-overlays`. Companion to the frontier surveys .run/{analyze_frontier,probe_hnorm}.py.
|
||||
"""
|
||||
import json, glob, re, collections
|
||||
import json, glob, os, re, sys, collections
|
||||
|
||||
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
||||
import corpus # the derived corpus oracle (Phase 26-A)
|
||||
|
||||
sig_by_ov = {}
|
||||
for p in sorted(glob.glob(".run/sig.ov_*.jsonl")):
|
||||
@@ -29,33 +32,50 @@ for p in sorted(glob.glob(".run/sig.ov_*.jsonl")):
|
||||
recs[int(d["addr"], 16)] = (d["nins"], d["h_exact"], d["h_norm"])
|
||||
sig_by_ov[ov] = recs
|
||||
|
||||
# matched set (h_exact): ov_SC01_077's non-stub functions + every dedup-registered hash
|
||||
src077 = "".join(open(cf).read() for cf in glob.glob("src/ov_SC01_077/*.c"))
|
||||
stub077 = set(int(m, 16) for m in re.findall(r'INCLUDE_ASM\([^)]*,\s*func_([0-9A-Fa-f]+)\)', src077))
|
||||
# ---- THE MATCHED SET, DERIVED (Phase 26-A audit, HIGH — R33) --------------------------------------
|
||||
# The old oracle asked ov_SC01_077 ALONE: matched := {h_exact of its non-stub fns} | dedup hashes.
|
||||
# That is the wrong question. A function absent from that one overlay — or stubbed there but matched
|
||||
# in the 133 others — came out "unmatched", so it was ranked as live work.
|
||||
#
|
||||
# MEASURED CONSEQUENCE: docs/family-manifest.md, the document the whole Phase-25/26 endgame is planned
|
||||
# from, advertised "2,758 multi-member families / 11.0 MB of hidden leverage". 1,071 of them (6.80 MB,
|
||||
# **62% of the advertised byte-weight**) were ALREADY FULLY MATCHED — pure phantom targets. A further
|
||||
# 253 had their byte-weight inflated by already-matched members. The byte-weight RANKING — the entire
|
||||
# purpose of the file, "draft these first" — was therefore sorted mostly on dead work, with the real
|
||||
# targets buried underneath.
|
||||
#
|
||||
# The invariant answers it exactly, per overlay, with no oracle at all:
|
||||
# an h_exact class is WORK iff at least ONE of its instances is still an INCLUDE_ASM stub somewhere.
|
||||
# (And the dedup-hash union is now redundant: a dedup-shared member is by definition not a stub. One
|
||||
# more scanner deleted — the `hash:` regex over config/dedup.us.yaml is gone.)
|
||||
stub_by_ov = {ov: set(corpus.stubs(ov)) for ov in sig_by_ov}
|
||||
sig077 = sig_by_ov.get("ov_SC01_077", {})
|
||||
matched_hex = {hx for v, (n, hx, hn) in sig077.items() if v not in stub077}
|
||||
matched_hex |= set(re.findall(r'hash:\s*([0-9a-f]{40})', open("config/dedup.us.yaml").read()))
|
||||
stub077 = stub_by_ov.get("ov_SC01_077", set())
|
||||
|
||||
# per h_exact class: {ovs, nins, rep_vram (ov_SC01_077's if present), h_norm}
|
||||
# per h_exact class: which overlays carry it, and in which is it STILL A STUB
|
||||
hexcls = {}
|
||||
for ov, recs in sig_by_ov.items():
|
||||
st = stub_by_ov[ov]
|
||||
for v, (n, hx, hn) in recs.items():
|
||||
e = hexcls.setdefault(hx, [set(), n, v, hn])
|
||||
e[0].add(ov)
|
||||
e = hexcls.setdefault(hx, {"ovs": set(), "stub_ovs": set(), "n": n, "rep": v, "hn": hn})
|
||||
e["ovs"].add(ov)
|
||||
if v in st:
|
||||
e["stub_ovs"].add(ov)
|
||||
if ov == "ov_SC01_077":
|
||||
e[2] = v
|
||||
e["rep"] = v
|
||||
|
||||
# UNMATCHED h_exact classes -> group by h_norm (the structural family)
|
||||
fam = collections.defaultdict(lambda: {"hexclasses": 0, "instances": 0, "nins": 0, "bw": 0, "rep": 0, "maxhexreach": 0})
|
||||
for hx, (ovs, n, v, hn) in hexcls.items():
|
||||
if hx in matched_hex:
|
||||
continue
|
||||
r = len(ovs)
|
||||
for hx, e in hexcls.items():
|
||||
if not e["stub_ovs"]:
|
||||
continue # fully matched fleet-wide -> NOT work (was: counted as work)
|
||||
n, hn, v = e["n"], e["hn"], e["rep"]
|
||||
r = len(e["stub_ovs"]) # the REAL x-N leverage: only the members still to bank
|
||||
f = fam[hn]
|
||||
f["hexclasses"] += 1
|
||||
f["instances"] += r # fleet-wide member count = the x-N propagation leverage
|
||||
f["instances"] += r
|
||||
f["bw"] += r * n * 4
|
||||
f["maxhexreach"] = max(f["maxhexreach"], r)
|
||||
f["maxhexreach"] = max(f["maxhexreach"], len(e["ovs"]))
|
||||
if n >= f["nins"] or f["rep"] == 0:
|
||||
f["rep"] = v
|
||||
f["nins"] = n
|
||||
@@ -63,7 +83,7 @@ for hx, (ovs, n, v, hn) in hexcls.items():
|
||||
# the draftable exemplar per family: the largest UNMATCHED ov_SC01_077 stub whose h_norm == the family
|
||||
ov077_ex = {} # h_norm -> (vram, nins)
|
||||
for v, (n, hx, hn) in sig077.items():
|
||||
if v not in stub077 or hx in matched_hex:
|
||||
if v not in stub077:
|
||||
continue
|
||||
if hn not in ov077_ex or n > ov077_ex[hn][1]:
|
||||
ov077_ex[hn] = (v, n)
|
||||
@@ -95,7 +115,7 @@ multi = [r for r in rows if r["multi_member"]]
|
||||
with_ex = [r for r in multi if r["ov077_exemplar"]]
|
||||
|
||||
# classify the multi-member families by ov_SC01_077 membership -> the endgame's 3 levers
|
||||
ov077_matched_hn = {hn for v, (n, hx, hn) in sig077.items() if v not in stub077 or hx in matched_hex}
|
||||
ov077_matched_hn = {hn for v, (n, hx, hn) in sig077.items() if v not in stub077}
|
||||
lev = {"draftable": [0, 0], "matched_free": [0, 0], "absent": [0, 0]}
|
||||
for r in multi:
|
||||
k = "draftable" if r["ov077_exemplar"] else ("matched_free" if r["hnorm"] in ov077_matched_hn else "absent")
|
||||
|
||||
Reference in New Issue
Block a user