Skip to content

feat: prepare better support for VMware virtual machines - #892

Merged
hivecommons-hive[bot] merged 15 commits into
mainfrom
hive/adopt-812
Nov 6, 2026
Merged

hivecommons-hive[bot] merged 15 commits into
mainfrom
hive/adopt-812

Conversation

@hanthor

@hanthor hanthor commented Nov 6, 2026

Copy link
Copy Markdown
Member

Summary: This prepares packages that help Utah work inside VMware virtual machines, including communication with the host and desktop integration. The archive-reading library those tools need is already on main (#846). The build instructions and verified source downloads are ready for review, but successful builds and working guest integration have not yet been demonstrated.

What I need from you: Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.

Re-homed from #812 by a maintainer: the fork could not be updated by the hive (fork guard / workflow permission). Original author: @mendezr.


Refs #488 — partial recipe work only; hosted builds, publication, permitted-repository installation and runtime acceptance remain outstanding. Consumer tracking: projectbluefin/utah#392.

Change

  • Import open-vm-tools 13.1.0 from Fedora Rawhide commit fde796d78f05621fce6852be53b70558ddb7f14e. Its unconditional desktop subpackage covers open-vm-tools-desktop; there is no duplicate source recipe.
  • libmspack (supplies pkgconfig(libmspack)) is already a published factory recipe from feat: add build instructions for the Microsoft CAB archive library #846. This PR does not touch it.
  • Add one HTTPS source lock (open-vm-tools, SHA-512 07086e24…90bbb4, matching the imported sources file) and one .packit.yaml entry. Move open-vm-tools and open-vm-tools-desktop to [resolved] with pr = 812.
  • Keep the imported patches, services, PAM file and run-vmblock\x2dfuse.mount unit. The recipe inherits no %check; its comment explains the documentation-font regeneration problem.
  • Update the factory-build-backlog skill. It records the source-family relationship, the existing libmspack (feat: add build instructions for the Microsoft CAB archive library #846), the build-root providers below, and guest-only runtime checks.

Build-root providers (answers @hanthor's 10-30 review)

None of these needs a recipe here. I checked each one against the primary metadata of every repository the mock build root pairs (tools/mock_config.py), fetched 2026-11-06:

BuildRequires Provider
xmlsec1-openssl-devel Hummingbird xmlsec1-openssl-devel-1.3.11-2.hum1; Fedora 44 1.2.41-4.fc44
pkgconfig(libtirpc) (libtirpc-devel) Hummingbird libtirpc-devel-1.3.8-0.hum1; Fedora 44 1.3.7-2.fc44
rpcgen Fedora 44 rpcgen-1.4-19.fc44 (Hummingbird has none)
procps-devel procps-ng-devel provides it: Hummingbird 4.0.6-2.hum1, Fedora 44 4.0.6-1.fc44
pkgconfig(gtkmm-4.0) / pkgconfig(libmspack) Factory gtkmm4.0-devel-4.24.0-1.hum1.bfin / libmspack-devel-0.11-0.1.alpha.hum1.bfin

The runtime xmlsec1-openssl and libtirpc also come from Hummingbird.

The published factory image (ghcr.io/projectbluefin/utah-packages@sha256:88d7a55ecfb729ab6f6944c0e8ebd282fe777ba7b59d1b44c7cf1a5997892d2f, run 41780828846) already ships packages that build against the same providers:

  • quota and nfs-utils (rpcgen, libtirpc-devel)
  • oath-toolkit (xmlsec1-openssl-devel)
  • igt-gpu-tools (pkgconfig(libproc2), also from procps-ng-devel)

Not proven yet: Hummingbird's priority will select xmlsec1 1.3.x over Fedora's 1.2.x, and whether VGAuth compiles against it is only shown by a build. That is the Canary on this PR. Whether the run-vmblock\x2dfuse.mount source is staged correctly is also left to that build.

Conflict resolution against main (0930e3b8)

This follows the scanner's 11-01 resolution plan on #812:

  • config/upstream-sources.json: kept main's list and added only the open-vm-tools entry at its sorted position. main now enforces name order (tools/source_lock_order.py).
  • docs/architecture.md: took main, which no longer quotes inventory counts.
  • reports/factory-build-backlog.json: deleted on main (fix: stop committing derived inventory counts and backlog totals #858), so it is not restored.
  • Dropped the obsolete fix: refresh inventory after main merges commit (db0fd8b2), because main no longer commits inventory. Its still-valid VMware note is folded into the new skill commit.
  • The .packit.yaml, backlog TOML and skill-doc VMware additions are kept.
  • mendezr's commit keeps its authorship and now carries a maintainer sign-off.

Validation

  • just check passed: validated 714 source RPMs (6 master, 702 rawhide, 6 upstream), with 1485 tests passed.
  • python3 tools/factory_build_backlog.py --check passed.

Remaining acceptance

No binary NEVRA, publication digest, installation transaction or VMware payload test is claimed. After the build:

  • Inspect the built RPM requirements.
  • Resolve installation using only Hummingbird plus the factory.
  • In a VMware guest, verify vmtoolsd, vmware-toolbox-cmd stat hosttime, desktop resizing and clipboard transfer.

mendezr and others added 2 commits November 6, 2026 02:03
Import the open-vm-tools source family and its required libmspack library. Lock upstream release archives, register both recipes and reconcile the three audit names. Keep hosted build, publication and guest runtime acceptance outstanding on #488.

Assisted-by: openai-codex/gpt-6.1-sol via GitHub Copilot
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Hive-Run: #488
Hive-Plan: VMware guest integration source-family import
Hive-Spec: issue-488#implementation-3
Signed-off-by: mendezr <mendezr@users.noreply.github.com>
Signed-off-by: James Reilly <jreilly1821@gmail.com>
… providers

Re-homing #812 onto main: libmspack is already a published factory recipe
(#846), so the skill no longer tells an importer to bring it in. Answer the
open review question on #812: Hummingbird provides xmlsec1-openssl-devel,
libtirpc-devel and procps-ng-devel (procps-devel), Fedora 44 provides rpcgen,
and the factory provides gtkmm4.0-devel, so no provider recipe is needed.

Assisted-by: Claude Opus via omp
Signed-off-by: James Reilly <jreilly1821@gmail.com>
@hivecommons-hive hivecommons-hive Bot added author/clanker A contributor's own hive authored this PR under their login dco All commits are DCO signed-off role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 6, 2026

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: This change adds the recipe for VMware guest tools, copied faithfully from Fedora. The recipe files themselves check out, but the description says the automated "Canary" check on this pull request is the requested build proof, and it is not: that check built nothing here. Two smaller wording gaps also need a fix so the project's records stay consistent.

What I need from you: Reword the proof claim and fix the two record inconsistencies below (suggestions included).


intent-alignment (medium)

The PR body says "its own Canary is the scoped hosted build the original review asked for" and "That is the Canary on this PR." The Canary run on head 2cf602c0 (job 127681269064) executed only the Decide step and printed no pipeline path changed; nothing to prove; every pass* job was skipped. That is by design: tools/canary.py:38-41 ("Recipes themselves are not here: they are proven by building them") and PIPELINE_PATHS exclude packages/. The repository's own guidance agrees — docs/skills/factory-build-backlog.md:770: "A green inventory check or Canary with skipped builds supplies neither proof". So no build of open-vm-tools exists for this head, and the xmlsec1 1.3.x / VGAuth question the body defers to "the Canary" remains unanswered. The scoped hosted build is rebuild-rpms.yml via workflow_dispatch with packages: ["open-vm-tools"]. Suggested body wording:

-**What I need from you:** Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.
+**What I need from you:** Review the import, then dispatch `rebuild-rpms.yml` with `packages: ["open-vm-tools"]` for the hosted build; the pull-request Canary does not build recipes.

docs-currency (low)

docs/skills/factory-build-backlog.md:347 (added) states libmspack "is already a factory recipe (#846, published)", while config/factory-build-backlog.toml at this head still records it as # Microsoft CAB compression library recipe import (#502); hosted build, publication and runtime proof remain outstanding. / { name = "libmspack", pr = 846 } with no published_run/published_nevra (compare the gnome-rounded-blur entry). The two files disagree at the same commit. Either drop "published" from the skill text, or — if the factory image in the PR body really ships libmspack-devel-0.11-0.1.alpha.hum1.bfin — update the libmspack backlog entry with its published_run/published_nevra and the comment. Minimal fix:

`open-vm-tools`; import the source family once. Its CAB library `libmspack`
is already a factory recipe (#846); do not import a second copy. Its

style (low, advisory)

config/factory-build-backlog.toml:392-393: the new [resolved] entries carry pr = 812, a closed, unmerged fork PR. Neighbouring entries (libmspack 846, libpaper 853) point at the PR that landed the recipe. Use pr = 892 so the provenance link resolves to merged history.

Limits: I verified open-vm-tools.spec, all three patches, both .service files, vmtoolsd.pam, open-vm-tools.conf and sources are byte-identical to Fedora rawhide fde796d7; the run-vmblock\x2dfuse.mount unit was read but not byte-compared. No build was attempted; a GitHub API rate cap ended further verification.

No findings from: correctness, security.

Confidence: 3/5 (needs attention) — 1 medium finding; 2 perspectives requested changes

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot added stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act and removed stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 6, 2026
@hivecommons-hive

Copy link
Copy Markdown
Contributor

@hanthor this PR is back in your court (stage/072-returned-to-contributor). The hive will not merge it until the head changes.

Review asks:

  • the review swarm requested changes

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.

@hivecommons-hive hivecommons-hive Bot added stage/050-review-pending Review swarm requested; awaiting a verdict for the current head and removed stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act labels Nov 6, 2026

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: This is a re-check of the VMware guest-tools import after its latest update. The only change since the last review is a routine sync with the main branch (a note about an unrelated text library being published); nothing about the VMware packages themselves moved. Two earlier concerns remain: the description still says the automated build on this branch proves the recipe builds, but that build did not compile anything, and the guide still says a helper library is published while the tracking list says it is not.

What I need from you: Fix the two wording points below (PR description and one doc line), or point to the evidence that makes them true.


Reviewed head 8dccf9c (delta from 2cf602c: merge of main #894, config/factory-build-backlog.toml +2/-1 and docs/skills/factory-build-backlog.md +4/-0, both slang-only). The delta is clean; previous findings re-checked below.

intent-alignment — still present (medium, in-scope)
PR body: "This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for." At this head the Canary job (run 42156575756, job 127692116366, head_sha 8dccf9c…) ran only tools/canary.py pr-relevant and exited with no pipeline path changed; nothing to prove; its CANARY_SET is ["libical", "vulkan-headers", "vulkan-loader"], so it never builds open-vm-tools on any head. No hosted build of this recipe exists yet. Suggested body edit:

-**What I need from you:** Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.
+**What I need from you:** Review the import. The PR Canary only builds its fixed set (libical, vulkan-*), so no hosted build of open-vm-tools has run yet; a scoped factory build is still needed before any build claim.

and drop "That is the Canary on this PR." under Not proven yet.

docs-currency — still present (low, in-scope)
docs/skills/factory-build-backlog.md:346-347 says libmspack "is already a factory recipe (#846, published)"; config/factory-build-backlog.toml:390-391 at the same head says "hosted build, publication and runtime proof remain outstanding" for libmspack. One of the two is wrong; since this PR owns the doc line and not the libmspack backlog entry, make the doc defer to the backlog:

`open-vm-tools`; import the source family once. Its CAB library `libmspack`
is already a factory recipe (#846; see the backlog for its publication state); do not import a second copy. Its

Previous style note (config/factory-build-backlog.toml:393-394 cite pr = 812 rather than this PR) is unchanged and remains advisory.

No findings from: correctness, security, style.

Confidence: 3/5 (needs attention) — 1 medium finding; 2 perspectives requested changes

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot added stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act and removed stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 6, 2026
@hivecommons-hive

Copy link
Copy Markdown
Contributor

@hanthor this PR is back in your court (stage/072-returned-to-contributor). The hive will not merge it until the head changes.

Review asks:

  • the review swarm requested changes

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.

@hivecommons-hive hivecommons-hive Bot added stage/050-review-pending Review swarm requested; awaiting a verdict for the current head and removed stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act labels Nov 6, 2026

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: Re-review of the VMware guest-tools import after the latest merges from the main branch. Nothing the author wrote changed since the last look; only merges from the main branch came in, and those are fine. The two notes from the previous review are still open: the skill document calls the archive library "published" when the project's own tracking file at the same revision says it is not yet published, and the description still calls the Canary check the build proof even though the Canary skipped building on this revision.

What I need from you: Fix the one doc word and reword the Canary claim (suggestions below).


Reviewed head d2b5340e (the dispatched revision); the PR has since moved to 9537f2e6 via another main merge, and the PR-authored hunks are byte-identical at 8dccf9c5, d2b5340e and 9537f2e6 (only hunk offsets differ).

Delta 8dccf9c5...d2b5340e: 87 files, all from main merges (ibus-*, libreport, satyr, #494/#804/#895). No PR-authored change. Nothing new to report.

Previous findings

docs-currency — still present (in-scope)
docs/skills/factory-build-backlog.md:346-347 says libmspack "is already a factory recipe (#846, published)", but config/factory-build-backlog.toml:390-391 at the same head reads "hosted build, publication and runtime proof remain outstanding" for libmspack. Same wording appears in the PR body ("already a published factory recipe from #846").

`open-vm-tools`; import the source family once. Its CAB library `libmspack`
is already a factory recipe (#846, build and publication still outstanding); do not import a second copy. Its

intent-alignment — still present (in-scope)
PR body: "its own Canary is the scoped hosted build the original review asked for." At d2b5340e the Canary job (run 42166425545, job 127723924821, head_sha confirmed) ended with no pipeline path changed; nothing to prove; all pass*/verify-* jobs were skipped. Per tools/canary.py:41-47,80-85, config/upstream-sources.json and config/factory-build-backlog.toml are inventory DATA_FILES, and packages/ is not a pipeline path, so this PR cannot trigger a Canary build of open-vm-tools. The body's "That is the Canary on this PR" (xmlsec1 / run-vmblock\x2dfuse.mount proof) is therefore not provided by this PR's CI. Reword the body to say the hosted build happens post-merge via the factory (as the "Not proven yet" paragraph already implies), e.g.:

-**What I need from you:** Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.
+**What I need from you:** Review the import. The Canary on this PR does not build recipes (only inventory files changed); the first hosted build of `open-vm-tools` happens in the factory after merge, and #488 tracks that proof.

No findings from: correctness, security, style.

Confidence: 3/5 (needs attention) — 1 medium finding; 2 perspectives requested changes

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: This is a re-check of the VMware guest-tools import after the branch picked up the latest main. Nothing in the import itself changed, so the two earlier notes still stand: the skill guide says the archive library the tools need is already published when the project's own backlog says it is not, and the description still points at an automated build that did not actually build anything. Both are small text fixes for the author.

What I need from you: @hanthor — apply the two wording fixes below (or confirm libmspack really is published and update the backlog instead).


Reviewed head 6f6974cd (compare d2b5340e...6f6974cd is four commits, all Merge branch 'main' plus the main-side #899/#903 content; 656508fe...6f6974cd is still exactly the 17 open-vm-tools files). The branch has since moved to d8e67917, which is another main merge (#906) and does not change either finding.

docs-currency — still present (in-scope)
docs/skills/factory-build-backlog.md:347 says libmspack is (#846, published), but config/factory-build-backlog.toml:383-384 at the same head keeps libmspack with no published_run/published_nevra and the comment "hosted build, publication and runtime proof remain outstanding". #906 (now merged into the branch at d8e67917) recorded publication for other packages and still left libmspack unpublished. The doc and the backlog disagree inside one PR.

is already a factory recipe (#846; hosted build and publication still outstanding); do not import a second copy. Its

intent-alignment — still present (in-scope)
The PR body says "its own Canary is the scoped hosted build the original review asked for." At head 6f6974cd the Canary job (run 42176475105, job 127756095505, head_sha = 6f6974cd) ran only the Decide step and printed no pipeline path changed; nothing to prove; pass1–pass6 and every verify job were skipped. That is by design — tools/canary.py:41-47 only compiles when a PIPELINE_PATHS entry changes, and tools/canary.py:74-79 explicitly excludes new recipe entries in config/upstream-sources.json. So no hosted build of open-vm-tools exists yet. Please reword the body, e.g.:

-**What I need from you:** Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.
+**What I need from you:** Review the import. The pull-request Canary skips recipe-only changes (it built nothing here); a scoped hosted build still needs a `canary.yml` `workflow_dispatch` with `lane_packages: ["open-vm-tools"]`, or the regular factory run after merge.

No findings from: correctness, security, style.

Confidence: 3/5 (needs attention) — 1 medium finding; 2 perspectives requested changes

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot added the stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act label Nov 6, 2026
@hivecommons-hive

Copy link
Copy Markdown
Contributor

@hanthor this PR is back in your court (stage/072-returned-to-contributor). The hive will not merge it until the head changes.

Review asks:

  • the review swarm requested changes

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.

@hivecommons-hive hivecommons-hive Bot added stage/050-review-pending Review swarm requested; awaiting a verdict for the current head and removed stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act labels Nov 6, 2026

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: Re-review after the branch picked up the latest main (yesterday's partition-mapping package, #898). That merge is clean and changes nothing about the VMware guest-tools import itself. The two issues raised last time are still open: the skill document says the archive library this package depends on is already published when the project's own backlog says it is not, and the description says this pull request's build check proves the recipe builds, when that check deliberately skipped building anything.

What I need from you: Fix the wording in the skill document (suggestion below) and reword the build claim in the description.


Reviewed head 4dde0829. Delta since ed8d0385 is only the merge commit bringing in d5dd441a (#898, packages/device-mapper-multipath/*); the PR's own diff vs main is unchanged (17 files, all open-vm-tools plus the four shared config/doc files, each resolved as a pure addition).

docs-currency — still present (medium, in-scope)

docs/skills/factory-build-backlog.md:358-359 says libmspack is "already a factory recipe (#846, published)". At the same head, config/factory-build-backlog.toml:379-380 records libmspack with the comment "hosted build, publication and runtime proof remain outstanding", and the PR body says the same. The skill doc is operator guidance; it must not claim publication the backlog denies.

`open-vm-tools`; import the source family once. Its CAB library `libmspack`
is already a factory recipe (#846; its hosted build and publication are still
outstanding); do not import a second copy. Its

intent-alignment — still present (medium, in-scope)

PR body: "its own Canary is the scoped hosted build the original review asked for" and "That is the Canary on this PR". At this head the Canary job (run 42201992685, head_sha 4dde0829) decided RUN=false and printed "no pipeline path changed; nothing to prove"; every pass*/verify-* job was skipped. This is by design: .github/workflows/canary.yml:111 fixes CANARY_SET to libical/vulkan-headers/vulkan-loader, and tools/canary.py:123-135 (pipeline_relevant) treats new recipes and new lock entries as inventory, not pipeline. No hosted build of open-vm-tools has happened, so the xmlsec1 1.3.x / VGAuth and run-vmblock\x2dfuse.mount staging questions the body defers to "the Canary" remain unanswered. Suggested rewording for the body's "What I need from you" line and the "Not proven yet" paragraph:

-**What I need from you:** Review the import. This branch is in-repo, so its own Canary is the scoped hosted build the original review asked for.
+**What I need from you:** Review the import. No hosted build of this recipe has run yet: the Canary only builds its fixed set and skips recipe-only changes, so a scoped build (e.g. a lane dispatch naming `open-vm-tools`) is still needed before #488 can record a build result.
...
-That is the Canary on this PR. Whether the `run-vmblock\x2dfuse.mount` source is staged correctly is also left to that build.
+That requires a hosted build that has not run yet. Whether the `run-vmblock\x2dfuse.mount` source is staged correctly is also left to that build.

No findings from: correctness, security, style.

Confidence: 3/5 (needs attention) — 2 medium findings

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot added stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act and removed stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 6, 2026
@hivecommons-hive

Copy link
Copy Markdown
Contributor

@hanthor this PR is back in your court (stage/072-returned-to-contributor). The hive will not merge it until the head changes.

Review asks:

  • the review swarm requested changes

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.

@hivecommons-hive hivecommons-hive Bot added stage/050-review-pending Review swarm requested; awaiting a verdict for the current head and removed stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act labels Nov 6, 2026

@hivecommons-hive hivecommons-hive Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary: Re-review of the VMware guest-tools import after it picked up the latest main-branch changes. Those merges brought in nothing of their own (the change against main is still exactly the VMware recipe), and all checks pass. Two earlier notes still stand: the description says the hosted build on this PR proves the recipe, but that build ran nothing; and the skill guide says a helper library is already published when the project's own backlog says it is not.

What I need from you: Reword the PR body's build claim and fix the one doc line (suggestion below).


Reviewed head 5bb6c950 (delta from 4dde0829: two Merge branch 'main' commits for #901 and #907; PR-vs-main diff is open-vm-tools files only).

intent-alignment — still present (medium, in-scope)
PR body: "its own Canary is the scoped hosted build the original review asked for." At this head the Canary job (run 42212198770, job 127873406270, head_sha 5bb6c950…) printed no pipeline path changed; nothing to prove and ran no build; its Run actions/checkout step was skipped. tools/canary.py:38-46 — PIPELINE_PATHS covers only .github/, tools/, config/, Containerfile, and the comment says "Recipes themselves are not here: they are proven by building them." The Canary set is fixed (libical, vulkan-headers, vulkan-loader), so it can never build open-vm-tools. Fix: replace that sentence in the PR body with "No hosted build of this recipe has run; the Canary on this PR only guards pipeline paths and built nothing. The first factory run after merge is the build proof." and drop "That is the Canary on this PR" under Not proven yet.

docs-currency — still present (low, in-scope)
docs/skills/factory-build-backlog.md:369 says libmspack is "(#846, published)", but config/factory-build-backlog.toml:377-378 at the same head says "hosted build, publication and runtime proof remain outstanding" and the entry carries no published_run/published_nevra. The PR body repeats the "published" claim.

is already a factory recipe (#846; hosted build and publication still outstanding); do not import a second copy. Its

No findings from: correctness, security, style.

Confidence: 3/5 (needs attention) — 1 medium finding; 2 perspectives requested changes

— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88

@hivecommons-hive hivecommons-hive Bot added stage/095-pre-merge-soak Lane admitted; the pre-merge soak window is running before the merge is attempted stage/080-awaiting-merge-admission CI green, verdict approve, no modifiers; the automerge lane has not admitted it yet and removed stage/050-review-pending Review swarm requested; awaiting a verdict for the current head stage/095-pre-merge-soak Lane admitted; the pre-merge soak window is running before the merge is attempted stage/080-awaiting-merge-admission CI green, verdict approve, no modifiers; the automerge lane has not admitted it yet labels Nov 6, 2026
@hivecommons-hive
hivecommons-hive Bot merged commit 3214072 into main Nov 6, 2026
21 checks passed
@hivecommons-hive

Copy link
Copy Markdown
Contributor

Hive status — updated in place; edits do not notify. Full history is in the Activity Monitor.

Evidence bundle `#8925bb6c95` sha256:fd008ec0e4133c56ade587f5b4567ea991b1175cc54c96b7d8dfb427f54076d0 — download from the dashboard or `hivectl review evidence #892`.

@hivecommons-hive

Copy link
Copy Markdown
Contributor

Merged by hive contributor-lane auto-merge: CI green for ≥30m, reviewer verdict approve, @hanthor was auto-promoted after 5 clean merged PRs in this hive's repositories.

@hivecommons-hive hivecommons-hive Bot removed role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/095-pre-merge-soak Lane admitted; the pre-merge soak window is running before the merge is attempted labels Nov 6, 2026
hivecommons-hive Bot pushed a commit that referenced this pull request Nov 9, 2026
Fill published_run/published_nevra for all 17 #488 outputs observed in
ghcr.io/projectbluefin/utah-packages@sha256:11318e23d0f093c58fa8136e48ae7c8acae9e849c272152055473b707aff5d72
(run 42559053279, revision ff2affe): mtr, nano, nfs-client-utils,
nfs-common-utils, nfs-utils, nfsv3-client-utils, nfsv4-client-utils,
nilfs-utils, ntfs-3g-system-compression, nvme-cli, nvtop, oddjob,
oddjob-mkhomedir, open-vm-tools, open-vm-tools-desktop, pinentry,
pinentry-gnome3. Point nilfs-utils and open-vm-tools* at their merged
adoption PRs (#715, #892) instead of the closed proposals (#702, #812).
Record that the image's factory-state label marks mtr,
ntfs-3g-system-compression and open-vm-tools as failed at their current
input digest, so their published RPMs are retained older builds.
Runtime proof stays open on the issue.

Refs #488

Assisted-by: Claude Opus via omp
Hive-Run: #488
Hive-Plan: factory-build-backlog dev-tools reconciliation
Hive-Spec: issue-488#acceptance-publication-evidence

Signed-off-by: Karibhou <Karibhou@users.noreply.github.com>
Co-authored-by: Karibhou <Karibhou@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

author/clanker A contributor's own hive authored this PR under their login dco All commits are DCO signed-off

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants