Skip to content

BS-32: the hash recipes of content identity as source (ContentId) - #460

Closed
humuhumu33 wants to merge 2 commits into
release/v0.4from
claude/issue-274-content-id
Closed

humuhumu33 wants to merge 2 commits into
release/v0.4from
claude/issue-274-content-id

Conversation

@humuhumu33

Copy link
Copy Markdown
Contributor

Part of #274: the labeled hash recipes of content identity (BS-32, artifact::content_id) as language-1.2 source, with six roots of ContentId: treeDigest, sourceId, semanticId, buildId, attestationId, and pdfRecipeId. It is the second slice of the project manager.

Draft, on release/v0.4. It imports only the standard library (Numeral, Sha256), so it stacks on no other slice.

Design

Each root writes its recipe's domain, a NUL, and its frames (§21.1), then returns the 32 bytes of the message's SHA-256. A frame is:

  • the big-endian four-byte length of its label, then the label;
  • the big-endian eight-byte length of its payload, then the payload.

The semantic ID frames the pinned Lean toolchain. Each root admits exactly allocation and overflow, and none admits recursion. Choosing and normalizing the files a digest covers, and serializing the IR, the closure, and the attestation body, stay in the host.

The oracle is the existing public code of content_id.rs. No Rust outside the case changes.

Evidence

Measured on the normative host over release/v0.4 at cf76090d4:

Leg Runs Agreed
named vectors 17, reaching every frame of every recipe, an empty path, a 300-byte path, and a message across blocks all
committed projects' trees and source inputs up to 4 KB, the 26 committed build manifests, and 100 seeded runs of each root, evaluator 287 trees, 283 source, 102 semantic, 128 build, 102 attestation, 101 PDF recipe all
the same with trees up to 64 KB and 500 seeded runs of each root, pinned rustc 3,733 all
planted defects, each on its own named vector 15 all detected

The oracle reproduces the recorded build ID of each of the 26 committed build manifests from its source and semantic IDs. Lean verifies the module, and the eligibility report gives each root exactly allocation and overflow.

Gates at the head

cargo fmt --check, validate-model (479 IDs), validate-spec-links (479 rows), clippy on the workspace with -D warnings, the Intel lane plan test, and the 17 RP cases pass.

CI is expected red until the regeneration

The six packages are not committed yet, so conformance_bs_32 stops at the stack-bound step after the differentials above have agreed. The landing regeneration will commit them.

🤖 Generated with Claude Code

humuhumu33 and others added 2 commits November 10, 2026 02:50
…ecipes

The second slice of the project manager as source (#274), on release/v0.4:
the row (register 478 -> 479, the §31 table, the RP-07 check and its case),
the scenario, the §17.23 prose, the CHANGELOG entry, the README's claim, the
CI part and its Intel lane, and the case, conformance_bs_32, which holds
ContentId's six roots to the recipes of artifact::content_id on hand-built
inputs, on the committed projects' trees and source inputs, on the IDs of
the 26 committed build manifests, each of whose build ID the oracle
reproduces, and on seeded inputs, with 15 planted defects each named by its
vector. VERIFICATION.md records the first run: the evaluator agrees on every
run of each root, rustc on 3,733, and every plant is detected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
bootstrap/src/ContentId.lex.tex over the standard library's Sha256: the
frame of §21.1 and the tree digest, source, semantic, build, attestation,
and PDF recipe IDs. Lean verifies the module and each root is eligible with
exactly allocation and overflow. The entrypoint, the lock, the six export
roots, and the project README follow.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@humuhumu33

Copy link
Copy Markdown
Contributor Author

Landed in #480 (merge aaf845507 on release/v0.4), which carries this change with the BS-25..34 batch, regenerated together and verified by full gate run 43135541484. The branch is kept.

@humuhumu33 humuhumu33 closed this Nov 12, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant