Claims rational lifecycle SBF component mix — 2026-09-07

Evidence level: real SBF ProgramTest component mix on hbox. This record does not establish an all-eight checked release, a local-validator founding, a devnet deployment, or a live cohort.

Scope

The regression distinguishes a Product's actual coordinate count N from a receipt representation descriptor width K. The test begins with a Product of 258 coordinates and a receipt descriptor of width 2. It first submits a well-formed hostile descriptor of width 3, then submits the canonical width-2 request. It uses real Claims and Token-2022 ELFs plus seeded Core, Claims, and Rent components.

The hostile transaction must refuse as RationalLifecycleSbfErrorV2::Instruction (0x5210) and roll back every named protocol account. The accepted transaction must create a real Token-2022 receipt mint whose terminally burnable profile has zero supply. The test also compares the Core Market, Claims aggregate, and Rent credit against their pre-activation values, so zero-supply receipt creation cannot silently move those facts.

Exact source and red control

The isolated old source checkout was detached and clean at 7dd2b49b38e0de2191c1546dcc75c27cb4685426. Its complete-source bundle had SHA-256 1a04f41b1645013658142bbe95588ba48e33a10eee90bc8e83134b1c54487df2. The filtered serial command was:

SBF_OUT_DIR=<red-sbf-out> CARGO_TARGET_DIR=/tank/dregg-build/dclutch-codex-20260907/target \
CARGO_BUILD_JOBS=6 SWARM_MEM_MAX=32G swarm-build cargo test --locked \
  --manifest-path programs/dclutch-claims-sbf/program-test/rational-lifecycle/Cargo.toml \
  --test lifecycle sparse_receipt_lifecycle_keeps_product_n_separate_from_representation_k \
  -- --exact --nocapture --test-threads=1

It failed exactly at RationalLifecycleSbfErrorV2::Market (0x5214), before the accepted width-2 activation. The preserved log is /tank/dregg-build/dclutch-claims-red.0dQYPZ/red.log; it records one failed and five filtered tests.

Fixed Claims ELF and green runs

Claims was rebuilt from a separate clean detached full source checkout at 872522e07b12d78406731285c264ba9cf5c48005. Its source bundle had SHA-256 da1617d42754b94236f02da59723535741a00ccb4eff75f716cd86299fe5aa2e. The fresh Claims-only SBF build was run through swarm-build with SWARM_MEM_MAX=32G and CARGO_BUILD_JOBS=6; its strict scan found neither SBF frame diagnostic form nor an already loaded scheduler no-op marker.

The fixed Claims artifact SHA-256 was:

d4b15932f4eac16e88c82d24996e37db4ea3602faebcae39d15419ded2e3a059

It was mixed only with these unchanged controls and caller:

The sparse filtered run passed one test with five filtered. Its log, /tank/dregg-build/dclutch-claims-red.0dQYPZ/green-sparse.log, records the hostile 0x5210 refusal, then real Token-2022 InitializeMintCloseAuthority, permissioned-burn extension initialization, and InitializeMint2 for the accepted zero-supply mint.

The existing equal-width control was also run serially against this same component set:

a_strangers_lamport_cannot_block_a_prepaid_receipt_activation

It passed one test with five filtered. Its log, /tank/dregg-build/dclutch-claims-red.0dQYPZ/green-equal-width-control.log, keeps the expected underfunding refusal RationalLifecycleSbfErrorV2::Rent (0x5215) before the accepted Token-2022 mint initialization.

Limits

All run material remains under /tank/dregg-build/dclutch-claims-red.0dQYPZ, including source commit files, artifact hash manifests, build logs, red output, and separate green output. This was deliberately a Claims component mix: it does not substitute for an all-eight fresh checked gate or a final frame capture pair.