You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Exact bytes of the B4 law disposition manifest, published as an immutable artifact. One row per law, one disposition from the closed set of twelve in the global claim adjudication §4.
289 laws — 180 core, 106 services, 3 judge. TrueResidue = 0. Every claim in every law file has a nameable stronger seat. laws.rs does not merely shrink; it ends.
Disposition totals: GeneratedRegister 95, TestPakBehaviour 69, GatedDebt 38, CompileFail 29, DeleteDeadSurface 26, and the remainder across the other seven.
The 38 GatedDebt rows are blocked by a mint gate, not by a missing seat — their subject has no constructor outside #[cfg(test)], so the claim cannot reach the seat that owns it until a consumer exists. That distinction is exactly the one the current ledger cannot see, and it is the reason the drain is a sequence rather than a sweep.
What a reader must not conclude
Stated in §0 of the artifact and repeated here because these are the three misreadings that would do damage:
Not that the drain manifest was wrong about tautologies. PR Foundation Closure B1a: core-crate type residue #12 removed the ~30 it counted, so the core DeleteTautology column is zero. That is not "core has no trivial assertions" — several surviving laws still re-read a literal the same statement wrote, named in their reasons.
Not that TrueResidue = 0 means every claim already has a home. It means every claim has a nameable stronger seat. Thirty-eight cannot reach it today.
Not that a DeleteDeadSurface row is a recommendation to delete. The adjudication's words are "deleted or given a real consumer before they may remain public." All 26 become live TestPakBehaviour rows the moment a consumer lands.
The count moved, and the delta is not a correction
The adjudication says 290; the W3 drain manifest counted 183 + 104 + 3 = 290 at f1f1f08. Against this basis the population is 289 = 180 + 106 + 3. The difference is four deletions and three additions across PRs #12 and #10, enumerated row by row in §0 — not a recount of the same tree.
Why this is published rather than filed
The lifecycle design was once an ordinary file rewritten in place, and a full red-team ran against bytes that no longer existed anywhere. This manifest is the most consequential conclusion the drain analysis produced, and until now it existed only as a local file. Same failure shape, one layer up: a proof system whose own disposition ledger is a ghost.
Gists are never rewritten. A revision becomes a new artifact superseding this one by relation, so the lineage stays citable.
Provenance note
Three cells in §1's basis table originally carried absolute local worktree paths. They were reduced to the worktree directory name before publication — branch and head SHA, which are the load-bearing identifiers, are unchanged. The digest above is over the published bytes and was verified by re-fetching the gist, not by trusting the upload.
Standing
Not an implementation source and not an authority over merged code. B4's execution is gated behind B1c, B2 and B3; this document is what B4 executes against once those land.
Exact bytes of the B4 law disposition manifest, published as an immutable artifact. One row per law, one disposition from the closed set of twelve in the global claim adjudication §4.
4E670A71E89DC4C45852ABD279AB609CD4F49CDDFC4E4D1942E349115DBC4408src/laws.rsat24b3582(PR Foundation Closure B1a: core-crate type residue #12),macros/macroc/src/laws.rsandtestpak/src/laws.rsat6bfda8c(PR Foundation Closure B1b: the services' coupled seat, the empty relation, and the derived watch set #10). Both commits are ancestors ofmain, so the basis is permanently citable in public history.What it establishes
289 laws — 180 core, 106 services, 3 judge.
TrueResidue = 0. Every claim in every law file has a nameable stronger seat.laws.rsdoes not merely shrink; it ends.Disposition totals: GeneratedRegister 95, TestPakBehaviour 69, GatedDebt 38, CompileFail 29, DeleteDeadSurface 26, and the remainder across the other seven.
The 38
GatedDebtrows are blocked by a mint gate, not by a missing seat — their subject has no constructor outside#[cfg(test)], so the claim cannot reach the seat that owns it until a consumer exists. That distinction is exactly the one the current ledger cannot see, and it is the reason the drain is a sequence rather than a sweep.What a reader must not conclude
Stated in §0 of the artifact and repeated here because these are the three misreadings that would do damage:
DeleteTautologycolumn is zero. That is not "core has no trivial assertions" — several surviving laws still re-read a literal the same statement wrote, named in their reasons.TrueResidue = 0means every claim already has a home. It means every claim has a nameable stronger seat. Thirty-eight cannot reach it today.DeleteDeadSurfacerow is a recommendation to delete. The adjudication's words are "deleted or given a real consumer before they may remain public." All 26 become liveTestPakBehaviourrows the moment a consumer lands.The count moved, and the delta is not a correction
The adjudication says 290; the W3 drain manifest counted 183 + 104 + 3 = 290 at
f1f1f08. Against this basis the population is 289 = 180 + 106 + 3. The difference is four deletions and three additions across PRs #12 and #10, enumerated row by row in §0 — not a recount of the same tree.Why this is published rather than filed
The lifecycle design was once an ordinary file rewritten in place, and a full red-team ran against bytes that no longer existed anywhere. This manifest is the most consequential conclusion the drain analysis produced, and until now it existed only as a local file. Same failure shape, one layer up: a proof system whose own disposition ledger is a ghost.
Gists are never rewritten. A revision becomes a new artifact superseding this one by relation, so the lineage stays citable.
Provenance note
Three cells in §1's basis table originally carried absolute local worktree paths. They were reduced to the worktree directory name before publication — branch and head SHA, which are the load-bearing identifiers, are unchanged. The digest above is over the published bytes and was verified by re-fetching the gist, not by trusting the upload.
Standing
Not an implementation source and not an authority over merged code. B4's execution is gated behind B1c, B2 and B3; this document is what B4 executes against once those land.
🤖 Generated with Claude Code