# SR01 shared revocation evidence: local review artifacts

Baseline: draft-saez-xcim-04, experimental profile `xcim-consent-sync-01`.

`shared-revocations.xml` is the editorial source for the shared-evidence section.
`reference.py` and `test_reference.py` exercise a shared-history component with
real Ed25519 signatures and public deterministic test keys. `c01_reference.py`
is a byte-preserved copy of the -03 reference dependency. Its presence does not
turn this component into a complete SR01 verifier.

Run with Python and the pinned dependencies in `requirements.txt`:

```text
python reference.py
python test_reference.py
```

The schema is structural; the tests cover authenticated scope, chained history,
incremental acquisition, missing pages, checkpoint freshness, signatures,
admission, time bounds and rollback/conflict cases. Consult `test-report.json`
for the actual executed count. No test produces operational `CONSENT_CURRENT`.

This component omits integrated original-grant/query verification, live address
control, grant-relative lifecycle validation, actual blockchain verification,
disk persistence, distributed recovery and performance measurement. Its replica
holds acquisition candidates in memory; production code must separate untrusted
staging from the authenticated durable view, promote only complete validated
histories atomically, and retain independently authenticated restrictions.
The page transaction test does not establish crash durability. Stream bootstrap
requires full history; compact snapshots are not specified in this revision.

The SR01 profile preserves finite C01 grants and their signatures. It introduces
no new human consent action, indefinite duration, campaign API or unsubscribe UI.
Shared checkpoints retain a 300-second experimental exposure budget; they are
not five-minute recipient renewals. Historical grant blockchain verification and
withdrawal anchoring retain C01 semantics. Checkpoints are trusted attester
projections, not independently observed absence proofs or mandatory chain writes.

Neither fixture success nor draft publication demonstrates interoperability,
production scalability or an operational attester integration.