Skip to content

Prepare coordinated 0.7.0 release candidate - #86

Draft
Taz33m wants to merge 7 commits into
mainfrom
codex/prepare-0.7.0
Draft

Taz33m wants to merge 7 commits into
mainfrom
codex/prepare-0.7.0

Conversation

@Taz33m

@Taz33m Taz33m commented Sep 5, 2026 •

Copy link
Copy Markdown
Owner

Scope

Prepare both distributions as 0.7.0 on top of merged PR #84 (9059f171).
The version branch retains all reviewed release-code fixes, updates coordinated
version checks and version-bound native qualification IDs, and documents the
new captured-evidence and separate L3 book-replay surfaces.

Historical corpus bytes, regression controls, frozen experiment code/evidence,
and private artifacts are unchanged. The 13 research findings remain deferred
in #85; this PR does not close that issue.

Publication hold

This is a draft release-preparation PR, not a release announcement. Version
0.6.0 remains the public installation target. No tag, package upload, citation
date change, or supported-version change is included.

  • The project's external-adoption release gate still requires evidence or an
    explicit maintainer decision to revise the gate; an unanswered invitation
    and repository-owned campaigns do not satisfy it.
  • GitHub's Dependabot endpoint reports alerts disabled (HTTP 403), so alert
    review is unverified rather than recorded as zero alerts.
  • The pypi environment is confirmed tag-only (v*); both PyPI Trusted
    Publishers still need confirmation before publication.
  • Final public metadata, release date, and post-index installation proofs
    belong to the eventual coordinated publication, not this preparation PR.

Verification

Verified source: b0f2403f7f337c8703b6689689e6f85dd0971cb8.

  • All nine Actions jobs passed on this exact head: Python 3.10–3.13, native
    orderbook-rs, gocronx, PythonMatchingEngine, the Go CLOB, and Nautilus L3.
    PR Consolidate native qualification and harden evidence publication #84's previous checks were not substituted for these checks.
  • Local Python 3.13: 671 passed, 1 skipped, 82.39% coverage. Black, flake8,
    mypy, Bandit, compileall, pip check, and whitespace checks passed.
  • Both wheels and sdists passed privacy and Twine checks; both source archives
    rebuilt to identical logical wheel contents, and repeat direct wheel builds
    were byte-identical with the source commit's timestamp.
  • Clean ownership/uninstall, exact-version facade dependency, resolver-driven
    seven-command probes, and SHA-pinned public 0.5.0 migration passed.
  • Installed lightweight wheel: 3/3 fixed cases, 25/25 generated traces,
    5,000 events, and 10/10 semantic capabilities without NumPy or psutil.
  • Artifact-only proof: installed this wheel in a fresh environment, built Rust
    solely from its extracted sdist, detected the injected control at event 173,
    reduced it to five events, reproduced exactly, and passed the correct native
    engine on that regression. JSON/JUnit and 10/10 coverage were verified.
  • A separate advisory audit found no known vulnerabilities in the freshly
    resolved macOS/Python 3.13 runtime dependencies, NumPy 2.5.2 and psutil 7.2.2.
    It does not cover optional extras, native dependency graphs, or replace the
    unavailable repository-wide alert review.

Local wheel SHA-256 values (verification artifacts, not published packages):

  • Conformance: 926851f8b2887580a777a0bfd025b00fe1e347320549a29560f16b70be8f6a56
  • Simulator: cd04c02975e0843f400decc98e741c10fd9aa2355ffdeaf8ef64fe8dda60ef51

This PR remains draft pending the release decision and review. No Cubic review
has run on this draft; passing Actions checks do not claim review approval.

This branch has not been deployed

No deployments
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