Repository navigation
fix: fetch two Python packages from their publisher - #862
mrbobbytables wants to merge 22 commits into
Conversation
Move python-argcomplete and python-dasbus to their verified PyPI source archives. Keep the existing SHA-512 locks and Fedora URLs as fallbacks. Hive-Run: projectbluefin#42 Hive-Plan: supply-chain source-lock migration batch Hive-Spec: issue-42#upstream-release-provenance Assisted-by: Auto mode via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com>
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Canary triage: this is a maintainer-review prerequisite, not a source/build failure. Run 41243793653 passed The existing After reviewing the diff, a maintainer must run Canary for a fork pull request on 🐝 Hive Agent: |
|
Needs a decision, not a merge: check Head: verdict pending @cd8244b, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to Updated 2026-11-06 09:50 UTC. |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (3 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Needs a decision, not a merge: check Head: verdict pending @3394d69, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
Document the maintainer-run proof required for fork source-lock changes and why local checks or another fork run cannot replace approval. Assisted-by: Auto mode via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com>
|
Canary is blocked on maintainer approval, not a package build failure: the fork check reports no maintainer-run proof and skips all build passes. Pushed signed-off documentation commit 🐝 Hive Agent: |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Rechecked returned head 🐝 Hive Agent: |
|
Rechecked the returned head: Canary run https://github.com/projectbluefin/utah-packages/actions/runs/41245691958 fails only because no maintainer-run proof exists for After reviewing this exact head, a maintainer must run: gh workflow run canary-fork-pr.yml -R projectbluefin/utah-packages --ref main -f pr=862 -f head_sha=d9009db2a618846bedb0223e5a08cd89a518791dPlease route this to maintainer review/dispatch rather than returning it for another head change. Keep the required Canary gate intact. 🐝 Hive Agent: |
|
Needs a decision, not a merge: check Head: verdict pending @d9009db, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
There was a problem hiding this comment.
Summary: This change points two Python libraries at the files their own publisher ships instead of copies kept by Fedora, and I confirmed the publisher's files are byte-for-byte what the lock expects. The code and config changes are sound. The only gap is that the pull request description leaves out one of the documentation changes it makes.
intent-alignment
- low, in-scope — The PR body's "Changes" list names one doc addition (the PyPI lookup paragraph,
docs/skills/supply-chain-provenance.md:205-207), but the diff also adds an 18-line subsection "Fork Canary approval is not a source-integrity failure" (docs/skills/supply-chain-provenance.md:227-244). The text itself is accurate — it matches.github/workflows/canary-fork-pr.yml:11-13,40-41,101-102(inputspr_number/head_sha, force-push tocanary/pr-<N>) — and it is pertinent to this very PR (itsCanarycheck currently fails for exactly this reason). Suggested fix: add a bullet to the PR body so reviewers know the section is intentional, e.g. "Document how a maintainer supplies Canary proof for a fork PR viacanary-fork-pr.yml."
Verified independently: streaming both files.pythonhosted.org URLs (config/upstream-sources.json:4869,4894) yields SHA-512 d07756a8…fa43a0a and ccf575d2…3d8a873, matching the retained lock digests; PyPI's /pypi/<project>/<version>/json returns the same sdist URLs; key order matches the 201 sibling feed+fallback_urls entries; tools/validate.py:162-193 accepts the baseline shrink.
No findings from: correctness, security, style, docs-currency.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Needs a decision, not a merge: check Head: verdict pending @da7f38a, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
The returned head still has a maintainer-only Canary prerequisite, not a package/build failure. Run https://github.com/projectbluefin/utah-packages/actions/runs/41255578175 skipped all build passes and failed only because no maintainer-run proof exists for After reviewing this exact head, an upstream maintainer must run: gh workflow run canary-fork-pr.yml -R projectbluefin/utah-packages --ref main -f pr=862 -f head_sha=71496f4f91b7a2d3ae94e4ca11de898d29e83f06The workflow refreshes the required fork check once its Canary passes. Please route this to maintainer review/dispatch rather than another contributor head change. Each merge or push invalidates the prior proof. No new commit or push was made, and the existing DCO-signed commits and required Canary gate remain unchanged. The existing supply-chain skill already documents this exact blocker. 🐝 Hive Agent: |
|
Rechecked current head This return needs maintainer review/dispatch, not a contributor code fix. The existing source changes and required gate remain intact. No new commit or push was made: another head change cannot supply approval and would invalidate any proof for the previous head. The relevant learning already ships in this PR and is covered by After reviewing this exact head, a maintainer must run: gh workflow run canary-fork-pr.yml -R projectbluefin/utah-packages --ref main -f pr=862 -f head_sha=71496f4f91b7a2d3ae94e4ca11de898d29e83f06That workflow supplies the proof and refreshes the required PR check after success. Please route this to maintainer review rather than returning it for an artificial new head. 🐝 Hive Agent: |
|
Needs a decision, not a merge: check Head: verdict pending @71496f4, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Current head After reviewing this exact head, an upstream maintainer must run: gh workflow run canary-fork-pr.yml -R projectbluefin/utah-packages --ref main -f pr=862 -f head_sha=47e9a7aaf5c6f5abaae53b963b7ca0f142be3f9cThe workflow refreshes the required fork check after its Canary succeeds. Please route this to maintainer review/dispatch rather than another contributor head change: every push or main merge invalidates previous-head proof. No new commit or push was made; the source changes, DCO-signed contributor commits, and required gate remain intact. This prerequisite is already documented in the PR’s skill guidance and — hive: backend=copilot model=Auto 🐝 Hive Agent: |
|
Current head After reviewing this head, a maintainer must run: gh workflow run canary-fork-pr.yml -R projectbluefin/utah-packages --ref main -f pr=862 -f head_sha=47e9a7aaf5c6f5abaae53b963b7ca0f142be3f9cThe workflow supplies the proof and refreshes the required PR check after success. Please route this to maintainer review rather than require an artificial head change: another push cannot supply approval and invalidates any previous proof. The source changes and gate remain intact; this constraint is already documented in this PR and 🐝 Hive Agent: |
|
Needs a decision, not a merge: check Head: verdict pending @47e9a7a, 20/21 checks green, red: Canary, soak 0s of 30m0s. Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
There was a problem hiding this comment.
Read the diff against main and re-verified it rather than trusting the body:
- Downloaded both new primary URLs and streamed them through
sha512sum:argcomplete-3.6.3.tar.gz→d07756a8…fa43a0a,dasbus-1.7.tar.gz→ccf575d2…3d8a873. Both match the unchanged locks, so this is a transport-only change. fallback_urlsis a real field:tools/source_pipeline.py:337-351validates it and passes[url, *fallback_urls]tofetch_with_fallbacks, so the Fedora lookaside still serves if PyPI fails.- Hashed
files.pythonhosted.org/packages/<hh>/<hh>/<hash>/…primaries already exist in the lock (pyaudio), sotools/upstream_bump.pyprecedent is in place. - Removing the two names from
config/fedora-primary-sources.txtis exactly whatcheck_fedora_primary_sources(tools/validate.py:162) requires once the primary leaves Fedora; leaving them would trip the staleness check. - The new Canary section links
.github/workflows/canary-fork-pr.yml, which exists and takeshead_shaas described.
Looks correct to me. The only thing standing between this and merge is the fork-Canary proof the PR body itself documents, which a maintainer has to dispatch. I have not approved or merged anything.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
@mrbobbytables this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
Re-homed to #909 as an in-repo branch ( |
* fix: fetch two Python packages from their publisher Move python-argcomplete and python-dasbus to their verified PyPI source archives. Keep the existing SHA-512 locks and Fedora URLs as fallbacks. Hive-Run: #42 Hive-Plan: supply-chain source-lock migration batch Hive-Spec: issue-42#upstream-release-provenance Assisted-by: Auto mode via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com> * docs: explain fork Canary approval for source migrations Document the maintainer-run proof required for fork source-lock changes and why local checks or another fork run cannot replace approval. Assisted-by: Auto mode via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com> * docs(lane-validation): record re-homing a fork PR whose Canary proof cannot survive A fork pull request whose fork Canary proof is discarded by every merge from main (or that changes .github/) cannot stay green as a fork PR. Record the maintainer path used for #812, #877 and #862: cherry-pick the PR's own commits onto an in-repo hive/adopt-<N> branch with sign-off and open a new PR. Assisted-by: Claude Opus via omp Signed-off-by: James Reilly <jreilly1821@gmail.com> --------- Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com> Co-authored-by: mrbobbytables <mrbobbytables@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com>
…that change no pipeline path (#900) * fix(canary): keep a fork's Canary proof across clean merges of main A maintainer-run Canary proves exactly one fork head. The repository requires branches to be up to date, so the hive merges main into the fork branch and the new head loses the proof: on #862 Canary passed at 345c7e6, hivecommons-hive[bot] pushed 48929a6 (parents 345c7e6 and main's 0787856), the pull request's diff stayed the same, and Canary went red again. The fork's Canary now also accepts a passed copy on canary/pr-<N> at an earlier head when tools/fork_canary.py carry_problem finds nothing unreviewed in between: every commit added off main is a two-parent merge of main into the proven branch, each merge's tree is exactly what git merge-tree makes from its parents (no resolved conflict, no hidden edit), the pull request's own diff is unchanged hunk for hunk, it touches no .github/ path, and the base is main. Any other commit still needs a new review and dispatch. The refresh job quietly re-runs a moved head's failed Canary once, so a merge during the maintainer run also recovers; it never reopens or approves for that head. Assisted-by: Claude via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com> * fix(canary): carry a fork proof only when main changed no pipeline path Review of #900: a carried proof made Canary green on a head whose merged pipeline was never built. carry_problem now also requires that what came in from main between the proven head's merge base and the new head's touches no pipeline path. Only docs/, site/, tests/, .agents/, .claude/, top-level Markdown and recipes outside the canary set are exempt; .github/, tools/, Justfile, config/, locks/, Containerfile, .packit.yaml, .gitattributes, reports/, files directly under packages/ and the canary set's own recipes all refuse, with a notice telling the maintainer to dispatch 'Canary for a fork pull request' again with the new head. canary.yml passes CANARY_SET to the proof command for that. Assisted-by: Claude via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com> * docs(canary): scope the fork re-homing note to pipeline changes on main Signed-off-by: James Reilly <jreilly1821@gmail.com> --------- Signed-off-by: James Reilly <jreilly1821@gmail.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com>
Summary: Two Python libraries now download their build files from the Python package publisher instead of Fedora. Their existing file checks are unchanged, and Fedora remains available as a backup.
What I need from you: Review and merge it.
Changes
python-argcompleteandpython-dasbusto their PyPI source archives.config/fedora-primary-sources.txt.canary-fork-pr.yml.Verification
just checkpassed;just testpassed (1,418 passed, 2 skipped).pre-commit run --all-filesreports one unrelated existing failure: ShellCheck SC2129 in.github/workflows/rebuild-rpms.yml.Remaining work
python-typing-inspectionstays on the list because its fixed native Canary availability contract requires a separate source change and native-runner validation.enchant2stays on the list because its candidate GitHub release archive did not match the existing SHA-512.graphenestays on the list because its candidate GitHub release URL returned 404.kde-filesystem(teamnames) andkf5(macros.kf5) remain Fedora-primary: these are Fedora package metadata and RPM macros, not upstream release archives. Their URLs are already pinned to fixed Fedora commits.Refs #42 — this batch moves two locks; 43 Fedora-primary packages remain, so the issue stays open.
— hive: backend=copilot model=gpt-5.6-luna-utility
🐝 Hive Agent:
contributor| SHA:ec6bae17