Skip to content

fix(release): adopt the hardened RRT 0.3.0 workflow - #357

Merged
acgetchell merged 2 commits into
mainfrom
fix/353-rrt-release-adoption
Nov 19, 2026
Merged

acgetchell merged 2 commits into
mainfrom
fix/353-rrt-release-adoption

Conversation

@acgetchell

@acgetchell acgetchell commented Nov 19, 2026 •

Copy link
Copy Markdown
Owner

Summary

  • Adopt published RRT 0.3.0 across the lockfile, local release recipes, and isolated GitHub Actions jobs.
  • Bind GitHub publication to the reviewed acceptance, captured draft, and retained asset identities. Verify the captured release and exact crates.io version after upload.
  • Time native dependency preparation and acceptance verification separately in the scoped benchmark attachment job.
  • Refresh uv, Rumdl, Semgrep, and the workspace dependency locks.

Closes #353

Live v0.5.5 release validation is tracked in #356.

Validation

  • just security passed with zero dependency and secret findings.
  • just ci passed on macOS ARM64 with 1,141 Rust tests, 559 Python tests, doctests, benchmark checks, and examples.
  • just review completed against origin/main for all 18 files on 345d509. The first review's asset-replacement finding is fixed, with regression coverage for missing assets and replacement IDs.

Publication boundary

The final local draft observation checks each retained asset ID and checks digests when GitHub provides them. CodeRabbit also flagged the window between that observation and the publication request. The release procedure requires the draft and assets to remain unchanged during approval and publication. GitHub's release update API provides no asset-inventory precondition. Release procedure · GitHub release update API

Summary by CodeRabbit

  • Release Improvements
    • Release publication now verifies that the accepted release, tag, commit, and selected assets still match before publishing.
    • Release checks confirm that the release is visible, the required archive is present, and the package is available in the registry.
    • Release verification can use the captured release target for additional identity checks.
  • Documentation
    • Updated release guidance with verification timing, publication checks, and recovery steps.

- Pin the published supplier across locked and isolated release runtimes.
- Bind publication to the reviewed acceptance and captured GitHub draft.
- Use shared remote visibility checks for the exact release and crate version.
- Time scoped dependency setup separately from benchmark verification.
- Refresh uv, Rumdl, Semgrep, and the workspace dependency locks.

Closes #353
- Reject missing or replaced draft assets before freezing the release.
- Reuse the asset identity check during crates.io publication.
@acgetchell
acgetchell enabled auto-merge (squash) November 19, 2026 07:18
@coderabbitai

coderabbitai Bot commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

Explain Change →

No actionable comments were generated in the recent review.

Walkthrough

The release tooling adopts RRT 0.3.0 and uv 0.13.9. Candidate publication now validates the reviewed acceptance, captured release target, and selected assets. Release verification delegates visibility checks to RRT and can receive the captured target.

Changes: Release workflow

Layer / File(s) Summary
Supplier pins and release job setup
.github/workflows/*, dependency-sources.toml, justfile, pyproject.toml, docs/RELEASING.md, tooling/README.md, tests/tooling/test_release_baseline.py, tests/tooling/test_toolchain_integration.py
Release tooling pins RRT 0.3.0 and uv 0.13.9. The benchmark workflow runs dependency builds in a separate step from acceptance verification. The related tests and release guidance reflect these updates.
Captured-draft publication
tooling/release_candidate.py, justfile, tests/tooling/test_candidate_transfer.py, tests/tooling/test_release_context.py, tests/tooling/test_release_flow.py, tests/tooling/test_release_publishing.py
The publish operation seals the reviewed acceptance against the captured target and checks selected asset identities before publication. Tests cover successful publication, mismatches, and draft receipt capture.
Release visibility verification
justfile, .github/workflows/publish-crates.yml, docs/RELEASING.md, tooling/README.md, tests/tooling/test_release_publishing.py
The verification recipe delegates release and registry checks to RRT’s visibility command. It accepts an optional captured target, which the publishing workflow passes to verification.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Change: Bug fix · Severity of issue fixed: Medium

Sequence Diagram(s)

sequenceDiagram
  participant ReleasePublish
  participant CandidateCLI
  participant CandidateSeal
  participant GitHubRelease
  ReleasePublish->>CandidateCLI: invoke publish with owner, selection, and tag
  CandidateCLI->>CandidateSeal: seal reviewed acceptance and captured target
  CandidateSeal->>GitHubRelease: validate release and selected asset identities
  CandidateCLI->>GitHubRelease: publish reviewed release using selected target
Loading

Possibly related issues

  • Issue 333 tracks release pin alignment for a different release; this PR also updates the RRT pins across manifests, recipes, and workflows.
  • Issue 356 covers live validation and publication of v0.5.5; this PR implements the adopted release workflow, while that issue retains the hosted and publication validation.

Possibly related PRs

Suggested labels: github-actions, release, python

Merge Risk: Minimal · up to 345d5

This change moves release tooling to RRT 0.3.0 and ties publication to the reviewed acceptance, the captured draft and its assets. The supplied review evidence shows no concrete defect. Live release validation is tracked separately, so only normal pre-release checks remain.

Pre-merge checks: Passed 3 · Inconclusive 1

Failed Failed checks: 1 inconclusive

Check name Status Explanation Resolution
Linked Issues check Inconclusive Issue #353 is the only active direct link. Its coding requirements are the RRT 0.3.0 adoption, pin alignment, scoped release jobs, and tests. The reviewed summary shows these changes:
- RRT pinned to …
Provide independent reviewable evidence that the RRT 0.3.0 entries in Cargo.lock and uv.lock agree with pyproject.toml, dependency-sources.toml, and the workflow and justfile pins. Acceptable evidence includes the output of a lock-consisten…
Passed Passed checks: 3 passed

Check name Status Explanation
Description Check Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check Passed The title "fix(release): adopt the hardened RRT 0.3.0 workflow" matches the main change. The PR moves the release tooling from research-repo-tools 0.2.9 to 0.3.0 across the workflows, justfile, pyproj…
Out of Scope Changes check Passed The changes match the PR objectives and issue #353:
- RRT 0.3.0 adoption and pin changes across workflows, justfile, tooling README, RELEASING.md, and dependency-sources.toml.
- uv 0.13.9 and rumdl 0.…
Full details: Title check

Explanation

The title "fix(release): adopt the hardened RRT 0.3.0 workflow" matches the main change. The PR moves the release tooling from research-repo-tools 0.2.9 to 0.3.0 across the workflows, justfile, pyproject, and dependency-sources, and it changes the release publish, verify, and seal flow. The title is specific and short. It names the dependency and the release scope. It does not cover the candidate-publication and visibility-check details, which the title does not need to cover. The conventional-commit prefix is consistent with a fix to the release tooling.

Full details: Out of Scope Changes check

Explanation

The changes match the PR objectives and issue #353:

  • RRT 0.3.0 adoption and pin changes across workflows, justfile, tooling README, RELEASING.md, and dependency-sources.toml.
  • uv 0.13.9 and rumdl 0.3.5 version refreshes, with the Semgrep and workspace lock refreshes named in the objectives.
  • The rumdl per-file ignore for MD076 list spacing in pyproject.toml supports the rumdl refresh and affects only changelog files.
  • The release-candidate changes (seal, publish_draft, require_candidate_asset_identities) and the publish workflow change implement the stated goal of binding publication to the reviewed acceptance, captured draft, and retained asset identities.
  • The release-benchmarks step split implements the scoped timing requirement.
  • Tests cover the changed behavior and the removed generated-script path.

No change was found that lacks a connection to the PR objectives or the linked issue. The lock-file changes are excluded from this review, so this verdict does not cover their contents.

Full details: Linked Issues check

Explanation

Issue #353 is the only active direct link. Its coding requirements are the RRT 0.3.0 adoption, pin alignment, scoped release jobs, and tests. The reviewed summary shows these changes:

  • RRT pinned to 0.3.0 in pyproject.toml, dependency-sources.toml, ci.yml, prepare-release.yml, release-benchmarks.yml, and the justfile uvx runtime pins.
  • uv raised to 0.13.9 in pyproject.toml, prepare-release.yml, and release-benchmarks.yml.
  • release-verify delegates visibility to research-repo-tools release visibility, replacing the local adapter and generated script.
  • release-benchmarks.yml runs just release-dependency-builds as a separate step before just release-receive, which separates the timings.
  • publish_draft and PublicationTransfer carry the accepted candidate through seal and publication, and asset identities are checked.
  • Tests cover the publication wiring, the captured-target and replacement cases, and the visibility command.

The requirement to confirm that the manifest, lock, source declaration, and workflow and recipe pins agree depends on Cargo.lock and uv.lock. Both files match the ignored-file rule !**/*.lock and were excluded from review. Their RRT 0.3.0 entries cannot be checked from reviewable evidence. The PR reports a passing just ci run, but that report does not confirm lock contents. The live v0.5.5 validation is assigned to #356 and is not a requirement of this PR.

Resolution

Provide independent reviewable evidence that the RRT 0.3.0 entries in Cargo.lock and uv.lock agree with pyproject.toml, dependency-sources.toml, and the workflow and justfile pins. Acceptable evidence includes the output of a lock-consistency check, such as an uv lock --check result or a Cargo metadata resolution that reports research-repo-tools 0.3.0. Alternatively, a reviewer can confirm the lock entries directly. Once that evidence is available, the assessment can be re-run.

Finishing Touches

Generate unit tests (beta)

  • Commit to this branch
  • Create a new PR
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Recent review info

Run configuration

Commits

Reviewing files that changed from the base of the PR and between 08b97e6 and 345d509.

Files ignored due to path filters: 2

  • Cargo.lock is excluded by !**/*.lock
  • uv.lock is excluded by !**/*.lock
Files selected for processing: 16

  • .github/workflows/ci.yml
  • .github/workflows/prepare-release.yml
  • .github/workflows/publish-crates.yml
  • .github/workflows/release-benchmarks.yml
  • dependency-sources.toml
  • docs/RELEASING.md
  • justfile
  • pyproject.toml
  • tests/tooling/test_candidate_transfer.py
  • tests/tooling/test_release_baseline.py
  • tests/tooling/test_release_context.py
  • tests/tooling/test_release_flow.py
  • tests/tooling/test_release_publishing.py
  • tests/tooling/test_toolchain_integration.py
  • tooling/README.md
  • tooling/release_candidate.py

Included review availability: This review used your included allowance. 6 included reviews remain after this review. Your included PR review attempts over the past 7 days set your current allowance at 8 reviews per hour.

@codecov

codecov Bot commented Nov 19, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.33%. Comparing base (08b97e6) to head (345d509).
✅ All tests successful. No failed tests found.

Additional details and impacted files
@@           Coverage Diff           @@
##             main     #357   +/-   ##
=======================================
  Coverage   98.33%   98.33%           
=======================================
  Files          16       16           
  Lines        7756     7756           
=======================================
  Hits         7627     7627           
  Misses        129      129           
Flag Coverage Δ
unittests 98.33% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

@acgetchell
acgetchell merged commit b61b7d5 into main Nov 19, 2026
38 checks passed
@acgetchell
acgetchell deleted the fix/353-rrt-release-adoption branch November 19, 2026 08:45
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.

release: adopt the hardened RRT workflow for the next la-stack release

1 participant