Johnb/cyberplanetary-protocol
1
0
Fork 0
Code Issues Pull requests Projects Releases 45 Wiki Activity
45 releases 45 tags
  • vlocality-kit-1.0.4 9b384499ca

    Locality conformance kit 1.0.4 Stable

    Johnb released this 2026-10-19 03:42:36 +00:00 | 30 commits to main since this release

    Locality conformance kit 1.0.4: documentation only. Same requirements, tests and tooling as 1.0.3.

    • The zip README names the re-released editions: CP 0005 0.15.0, CP 0006 0.13.0, CP 0008 0.4.2, CP 0009 0.9.0, CP 0010 0.7.0 and CP 0011 0.3.2 (live on cyberplanetary.org since the PUBLIC108 publication; readback of 284 paths, no mismatch). The editions cite the kit as the range ^1.0.3, so 1.0.3 and 1.0.4 are both accepted; a conformance claim names the exact release used.
    • CP 0005 0.14.1, CP 0006 0.12.1, CP 0009 0.8.1 and CP 0010 0.6.1 were classed as patches although CP 0005 changed a requirement. Under CP 0000 they are replaced by the editions above and withdrawn, with a note. See CLASSIFICATION-CORRECTION.md in the evidence package.
    • Provider A (Python) 72/72, Provider B (Node.js) 72/72 and Client C (Python) 112/112 on the public suite and on the regenerated hidden instance, also from a clean unpack of these zips. Releases 1.0.0 and 1.0.1 remain superseded; 1.0.2 and 1.0.3 carry correction notes.

    Assets (sha256): locality-kit-1.0.4.zip c9563cefbe856db6fc5f7f99ad25b4052193b30a555e69e84645aea4ffb893dc, locality-evidence-1.0.4.zip 3781ac32e5224efb9ab439de8902dde977d209af91f68eac2e9e834fe57aede0, locality-terrain-1.0.0.zip 81b7b222f3528d46556c2c486d51fa1bfe20a4368dbd948a2dbdcc06b32df1eb

    Downloads
  • vlocality-kit-1.0.3 9115800232

    Locality conformance kit 1.0.3 Stable

    Johnb released this 2026-10-19 01:00:24 +00:00 | 32 commits to main since this release

    NOTE (kit 1.0.4): the standards editions now cite this kit as the range ^1.0.3 and name the re-released editions CP 0005 0.15.0, CP 0006 0.13.0, CP 0008 0.4.2, CP 0009 0.9.0, CP 0010 0.7.0 and CP 0011 0.3.2. CP 0005 0.14.1, CP 0006 0.12.1, CP 0009 0.8.1 and CP 0010 0.6.1 (mentioned below) are withdrawn. This release remains valid; see vlocality-kit-1.0.4.

    Locality conformance kit 1.0.3: corrects the hidden-suite mode and generator of 1.0.2. No contract, requirement, edition or candidate-facing change.

    Found by the independent review of 1.0.2 and reproduced here:

    • locality-assess --hidden DIR ran the PUBLIC suite and public fixture values: the public assessor directory preceded the hidden tree on the module path. Earlier "hidden suite" pass statements (kit 1.0.0 to 1.0.2) are withdrawn; they were public-suite passes. Hidden mode now loads the hidden tree, records the files it loaded and the fixture epoch in the report, and refuses to run otherwise.
    • The hidden-suite generator produced an unusable instance (renames reordered sorted lists, a grant key was missed, watertight vectors at the wrong level). Fixed; the generator and its output ship in the evidence package. C-TILE-02 compared the tile face with a literal instead of the suite parameter (fixed); the mock preload control now works against a hidden instance.
    • Results: selftest PASSED; Provider A (Python) 72/72, Provider B (Node.js) 72/72, Client C (Python) 112/112 on the public suite and, for the first time, on a genuinely different hidden instance (epoch 1835000000000, no module outside the hidden tree). The candidates were not changed.
    • The six standards editions (CP 0005 0.14.1, CP 0006 0.12.1, CP 0008 0.4.1, CP 0009 0.8.1, CP 0010 0.6.1, CP 0011 0.3.1) are unchanged; they name kit 1.0.2, whose public suite and conformance claims are unaffected. Use 1.0.3 for hidden-instance assessment.

    Assets (sha256): locality-kit-1.0.3.zip d93361d2d2c6a05d3e67b184772410b7201d0960981a189956395f45c957e855, locality-evidence-1.0.3.zip d81b0fc9a71a3b8fd7595bf4c026fee8b9fe19fd7247b54c83fa4df044b7e70e, locality-terrain-1.0.0.zip 81b7b222f3528d46556c2c486d51fa1bfe20a4368dbd948a2dbdcc06b32df1eb

    Downloads
  • vlocality-kit-1.0.2 9115800232

    Locality conformance kit 1.0.2 Stable

    Johnb released this 2026-10-18 16:24:32 +00:00 | 32 commits to main since this release

    NOTE (kit 1.0.4): the standards editions now cite this kit as the range ^1.0.3 and name the re-released editions CP 0005 0.15.0, CP 0006 0.13.0, CP 0008 0.4.2, CP 0009 0.9.0, CP 0010 0.7.0 and CP 0011 0.3.2. CP 0005 0.14.1, CP 0006 0.12.1, CP 0009 0.8.1 and CP 0010 0.6.1 (mentioned below) are withdrawn. This release remains valid; see vlocality-kit-1.0.4.

    CORRECTION (see vlocality-kit-1.0.3): the statements below that the candidates pass the "hidden" suites are withdrawn. In kit 1.0.0 to 1.0.2, locality-assess --hidden ran the public suite. The public suite, the editions that name this kit, and the issuer-identity and kit-closure repairs are unaffected. Use kit 1.0.3 for hidden-instance assessment.

    Locality conformance kit 1.0.2: corrects 1.0.0 and 1.0.1.

    Changes since 1.0.1, after the independent second review:

    • Issuer-authorisation identity is now (world, issuing context, id) in CP 0005 0.14.1 (L-RES-2 step 6, L-RES-2a(v), L-ISS-1). Two regional managers may choose the same issuer id; a client visiting both no longer reports a false conflicting-evidence. Contradictory evidence under the same qualified identity is still conflicting-evidence. The kit adds five chain cases (C-CHAIN-pos-issuer-same-id-two-contexts, -reverse, -generation-per-context, neg-equivocation-issuer-qualified, neg-fork-issuer-qualified), provider test P-Q-18 and two negative controls.
    • The kit is self-contained: the contract text, its changelog, CONTRACT-QUESTIONS.md and ORACLE-FINDINGS.md that the inner manifest names are now shipped; KIT/tools/coverage.py, check_vectors.py, gen_reqmap.py and locality-assess list run from a clean unpack (verified by the build).
    • Standards: CP 0005 0.14.1, CP 0006 0.12.1, CP 0008 0.4.1, CP 0009 0.8.1, CP 0010 0.6.1, CP 0011 0.3.1; CP 0005, 0008, 0009, 0010 and 0011 name this kit release (CP 0006 defines no kit reference). Releases 1.0.0 and 1.0.1 are superseded and must not be used.
    • Candidates (Python provider, Node.js provider, Python client) pass the public suites 72/72, 72/72, 112/112 and the hidden suites, from a clean unpack of locality-kit-1.0.2.zip.

    Assets (sha256): locality-kit-1.0.2.zip 5da6df9893b6e8da28015004e6fd310d00d754128b7738a4f6e4a8b22ac248e9, locality-terrain-1.0.0.zip 81b7b222f3528d46556c2c486d51fa1bfe20a4368dbd948a2dbdcc06b32df1eb

    Evidence package (added after the production publication): locality-evidence-1.0.2.zip 24ee66f9098c8065794263e26040590e313faca6023b6e300c575b0429d2d06a. It carries the hidden-suite generator and cases, the readback receipts of the PUBLIC107 publication (278 paths, 0 mismatches), and the response to the second review.

    Downloads
  • vlocality-kit-1.0.1 d46d953da2

    Locality conformance kit 1.0.1 (superseded) Stable

    Johnb released this 2026-10-18 13:58:25 +00:00 | 33 commits to main since this release

    SUPERSEDED: do not use. Use vlocality-kit-1.0.2 (issuer-authorisation identity corrected to world/context/id; self-contained kit).

    Locality conformance kit 1.0.1 (supersedes 1.0.0, which stays available unchanged), the terrain dataset 1.0.0 and the evidence package 1.0.1.

    • locality-kit-1.0.1.zip (sha256 1f0b611951db42b45afce72b032e471cf772704b243181a21dfd3a7cf8413491): the public implementer kit. New in 1.0.1: paired-case tests for private content versus public structural addition, proven-absent prohibition tests, an optional --selections declaration echoed in assessor reports, and the clarified contract text of draft 2.3. Candidates assessed against 1.0.1: provider in Python 71/71, provider in Node.js 71/71, client in Python 107/107 (public and hidden suites). A provider (Python) and the client were also re-run from this zip unpacked into an empty directory: both pass.
    • locality-terrain-1.0.0.zip (sha256 81b7b222f3528d46556c2c486d51fa1bfe20a4368dbd948a2dbdcc06b32df1eb): public terrain test dataset (complete coarse base levels 0-3, sparse finer tiles, contribution tiles for the two demonstration localities, exporter and verifier scripts, source-input hashes, provenance). Derived from a fictional planet model (radius 6,371,000 m); no tile is native source data.
    • locality-evidence-1.0.1.zip (sha256 343b16e2e0ee987acd35e1f16235459a9ce2a946343a5971a4895db10e805e58): implementations, assessment runs, scale experiment, multi-host demonstration including the outage rerun with the blind client, catalog registration, browser check and design records.
    Downloads
  • vlocality-kit-1.0.0 d46d953da2

    Locality conformance kit 1.0.0 (superseded) Stable

    Johnb released this 2026-10-18 02:07:34 +00:00 | 33 commits to main since this release

    SUPERSEDED: do not use. Use vlocality-kit-1.0.2 (issuer-authorisation identity corrected to world/context/id; self-contained kit).

    Locality conformance kit 1.0.0 and evidence package for the planetary-scale locality editions
    (CP 0005 0.14.0, CP 0006 0.12.0, CP 0008 0.4.0, CP 0009 0.8.0, CP 0010 0.6.0, CP 0011 0.3.0), published at repository commit d46d953.

    • locality-kit-1.0.0.zip (sha256 0380224e7ba24e7b5c029386c8b450624519483aa6a4e30d6fd392a7e973294b): the public implementer kit: standards closure, closed schemas, vectors, the runnable assessor (KIT/assessor/locality-assess, 172 tests), negative controls, public test assets (terrain tiles for two localities, expected manifests, content), and the maintained supporting roles (ROLES/: authority tooling, managers/issuer, base delivery, testbed). Hidden cases generated from the same public requirements are not distributed.
    • locality-evidence-1.0.0.zip (sha256 4db27c6c307b6fe57ced9acead3e50cbcff43e8694f1328534e429228bc15e62): the evidence package: the three independently built implementations (provider in Python, provider in Node.js 22, client in Python) with their assessor reports, question records and independence audit; the lead's hidden-suite runs; the multi-host demonstration report and raw traces; the metadata-scale experiment (1,000 / 10,000 / 100,000 synthetic parts, 1,000 principal fixtures); contract drafts and clarifications; terrain provenance.

    Claims: one world identity operated by 5 actual hosts with 3 independently built services (providers A and B, client C); authority, managers, issuer and base delivery are maintained reference roles and are not claimed as independent. The scale experiment is a 100,000-entry metadata exercise (9 service processes on one machine, repeated on a second), not 100,000 running simulators and not a client-load benchmark. Terrain data are derived from a fictional planet model (radius 6,371,000 m, the fixture's declared radius); no tile is native source data (labels: resampled/downsampled).

    Downloads
  • vconformance-kit-b-1.0.5 7e293cbab7

    Milestone B conformance kit 1.0.5 Stable

    Johnb released this 2026-10-15 20:18:43 +00:00 | 34 commits to main since this release

    Milestone B conformance kit 1.0.5 corrects the public assessment methods and their reproducibility records. It validates known supplied reference identities and realizations on unused paths while leaving genuinely absent unused dependencies nonblocking. The public raw auditor applies independent terminal action proof to submit, retry and status, preserving observation authentication checks.

    The positive manifest assurance fixture uses the delivered files layout. The exhaustive generator plan covers all 155 generated members, including the nine selective-reference cases and four indexes. Synthetic public controls cover complete nonvisual and geometric report comparison, canonical fixture identities and negative sensitivities.

    The exact selected core is cp-milestone-b-core/1 version 1.0.5, SHA-256 c9a086963a2f73bacda0029d30edab32447b1ddf024da02a8ebe4941d192e57e, with 3,791 committed members. CP 0005 v0.13.3, CP 0006 v0.11.3, CP 0009 v0.7.3 and CP 0010 v0.5.3 select these same bytes. CP 0011 v0.2.1 and existing wire meanings remain unchanged.

    Extract the binary kit and use its candidate/ root. Its README and reproduction/PUBLIC105-REPRODUCTION.json identify the complete verifier, public controls, generator replay and fresh handoff commands. compare_observation is the complete observation/report entry point used by the public trace auditor; an expected invalid-response predicate must come from independently selected fixture evidence.

    The kit contains no private candidate assessment or implementation. Fixed controls, preparation-only runs and support claims do not award implementation conformance. Previous releases retain their original bytes and method identities.

    Downloads
  • v0.13.3 7e293cbab7

    Reference source and standards 0.13.3 — public conformance method correction Stable

    Johnb released this 2026-10-15 20:18:26 +00:00 | 34 commits to main since this release

    Reference source and standards 0.13.3 publish the bounded Milestone B public-method correction. Four owning patch editions select conformance core 1.0.5, SHA-256 c9a086963a2f73bacda0029d30edab32447b1ddf024da02a8ebe4941d192e57e: CP 0005 v0.13.3, CP 0006 v0.11.3, CP 0009 v0.7.3 and CP 0010 v0.5.3, with publication bindings v0.13.3.

    The companion binary conformance kit 1.0.5 corrects known-reference consistency, terminal-only raw-audit applicability, the positive manifest control and complete generator commitments. All four editions select its exact 3,791-member core. CP 0011 v0.2.1, wire identifiers, simulation behavior and previously published editions remain unchanged.

    This publication changes conformance method selections and their companion source material. It does not deploy new worlds or products and grants no aggregate candidate conformance award. Assessment results retain their exact source, method and role scopes separately from the public kit.

    Downloads
  • vconformance-kit-b-1.0.4 0fcebb35a3

    Milestone B conformance kit 1.0.4 Stable

    Johnb released this 2026-10-15 11:51:11 +00:00 | 35 commits to main since this release

    Milestone B conformance kit 1.0.4

    The frozen PUBLIC kit supplies the complete candidate-root layout, fresh-input and reviewed-plan execution procedure, oracle, schemas, fixed cases and version-qualified owning standards. It retains all 142 original catalogue rows and adds nine requester selective-path cases. Candidate implementations, private assessment adapters and private qualification evidence are excluded.

    The spatial method validates all supplied structural meaning, then resolves the selected path's dependencies. An unrelated path's missing endpoint does not erase an otherwise valid selected placement. The startup rejection method preserves precise CP 0011 0.2.1 interpretation reasons while allowing bounded human-readable detail; it no longer demands that explanatory prose equal the canonical reason token. Positive, negative, compiled-input and raw-evidence checks remain substantive requirements.

    A successful development smoke is not a complete conformance-class award. Historical executions keep their actual input selections; qualification requires explicit dependency and role review.

    Core 1.0.4 SHA-256: d6698e20f20317d23ad9d1d04558da56de332f28dc47a8cbd5dfa630e9b361a0. Candidate inventory SHA-256: 0e77eec1397c57d8643d66bfbba51e66b485275d2739e437fa2027ad8064eaf8. Owning editions: CP 0005 0.13.2, CP 0006 0.11.2, CP 0009 0.7.2 and CP 0010 0.5.2. CP 0011 0.2.1 and accepted Milestone A are preserved.

    Archive: milestone-b-conformance-kit-1.0.4.tar.gz, SHA-256 efb8868ba44279b4e45260a902f995366eaa51a279a83652b8d0ffcd218a2674, 15,339,146 bytes. Source revision: 0fcebb35a3154d8b3b601f28e759a23f25ec53ec.

    Downloads
  • v0.13.2 0fcebb35a3

    Reference software 0.13.2 — Milestone B bounded correction Stable

    Johnb released this 2026-10-15 11:50:25 +00:00 | 35 commits to main since this release

    Reference software 0.13.2: Milestone B bounded correction

    Publishes coherent owning editions CP 0005 0.13.2, CP 0006 0.11.2, CP 0009 0.7.2 and CP 0010 0.5.2, with publication-only concept bindings 0.13.2. Their selected core 1.0.4 is SHA-256 d6698e20f20317d23ad9d1d04558da56de332f28dc47a8cbd5dfa630e9b361a0.

    The correction repairs selective spatial-evidence interpretation and the executable public candidate-root handoff. All supplied structural meaning is validated, while an unrelated path's unavailable endpoint does not erase a valid selected placement. The catalogue preserves all 142 original rows and adds nine requester selective-path cases.

    Startup configuration rejection retains the precise interpretation reason required by CP 0011 0.2.1 section 4.2. Its public method now permits bounded human-readable detail instead of requiring that prose to equal the reason token. Positive controls, mutation proofs, rejection classification and raw-evidence checks remain required. A successful development smoke does not award a conformance class.

    This is a standards and method release. The assessed runtime selections remain JavaScript 608c2509e8258ee4e6da18ac658a4dad67e44235 and independent Python 196d3ba4ccae2237900f6a7eabea822072ca78e0. Accepted Milestone A, CP 0011 0.2.1, wire identifiers and published history are preserved.

    Exact maintained source revision: 0fcebb35a3154d8b3b601f28e759a23f25ec53ec. Attached full source archive: cyberplanetary-protocol-0.13.2-source.tar.gz, SHA-256 86a52829112ea94f699a820d4bfeb782f088f1e89633a28089810166edbb31e3, 24,808,465 bytes. This release does not deploy production worlds.

    Downloads
  • vconformance-kit-b-1.0.2 2be1e1b660

    Conformance Kit B 1.0.2 Stable

    Johnb released this 2026-10-15 04:22:41 +00:00 | 37 commits to main since this release

    Complete implementation-neutral live producer/requester conformance kit1.0.2. This immutable method correction derives unavailable-authentication observations from assessor-owned transport evidence and includes explicit fixed controls rejecting discarded confirmation, accepted expired responses and discarded known records. The selected standards, schemas, corpus and candidate interface are unchanged.

    Publication supplies test methods and does not award candidate conformance or deployment availability. Prior1.0.1 bytes and scoped assessment evidence remain preserved.

    Downloads