Repository navigation
feat: add the virtual machine file-sharing service - #761
Conversation
f34c61e to
db9b193
Compare
There was a problem hiding this comment.
Summary: This review looked at the new build recipe for the virtual-machine file-sharing service. The recipe is a faithful copy of Fedora's, with two deliberate edits, and it fits how the other recipes here are laid out. I found nothing that should stop it from merging, but one of the new edits turns on an upstream test that opens a network port, and that test has never been shown to pass in a package build, so it may be the first thing to break when the real build runs. I also noted two small text tidy-ups.
Reviewed at head db9b193. Spec diffed against Fedora dist-git 6f5ddf3d: the only differences are the dropped || cat fallback and the added %check, as the PR says.
correctness
- low —
packages/phodav/phodav.spec:74-75adds%check/%meson_test. Neither Fedora's spec nor upstream's own CI (.gitlab-ci.ymlatv3.0builds, then runs litmus, nevermeson test) runs this test. The test (tests/virtual-dir.c:22,:106) spawnstests/virtual-dir-serverand talks to a hard-codeddav://localhost:8080/. It needs a working loopback interface and a free port 8080 inside the build root, and it will collide if two builds share a network namespace. libsoup3 does accept thedav://scheme for a direct connection (soup-session.conly gates scheme on redirects), so the test is plausible, just unproven. Nothing to change before the canary run; if it fails there, the remedy is to drop the%checkblock rather than patch the test.
security
- low —
packages/phodav/phodav.spec:9keeps Fedora's plain-http://Source0. The pipeline fetches from the HTTPS lock inconfig/upstream-sources.json, so this is metadata only, but the PR body says the archive is "locked over HTTPS" and the spec line contradicts it.Source0: https://download.gnome.org/sources/%{name}/%{version}/%{name}-%{version}.tar.xz
style
- low —
config/factory-build-backlog.toml:340is the only[resolved]entry without the one-line comment every neighbour carries (:341,:344,:346) saying what remains outstanding.# Recipe import only; build, publication, dependency closure and guest file-transfer proof remain on #490. { name = "spice-webdavd", pr = 761 }, - info —
docs/skills/factory-build-backlog.md:235-236introduces two consecutive blank lines before thespice-vdagentparagraph; drop one.
No findings from: intent-alignment, docs-currency.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Ready for a human merge: verdict approved @db9b193, 21/21 checks green, soak 30m0s elapsed. Blocked only by The hive never approves on your behalf: |
|
@thatguywob this PR is back in your court ( Review asks:
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 (2 waiting): fixes are handed out before new work as soon as you have capacity. |
|
@thatguywob the hive has adopted this PR ( |
hanthor
left a comment
There was a problem hiding this comment.
Changes requested (small supply-chain fix; runtime verification still needed). packages/phodav/phodav.spec:9 still declares Source0 as http://download.gnome.org/..., while the new lock uses the HTTPS URL. The locked pipeline verifies SHA-512, but a direct spec-driven fetch still uses unauthenticated HTTP. Please switch Source0 to HTTPS, retaining the pinned digest. Separately, the review does not establish package build/solver success or a representative SPICE/WebDAV guest transfer; keep the issue open until those are exercised. Reviewed with Claude.
|
The hive adopted this PR ( Blocker: lane reason Moving to |
|
hive/scanner: conflicts with main; not resolved by hive because PR lacks stage/074-adopted-by-hive — contributor please merge main. 🐝 Hive Agent: — hive: agent=scanner backend=copilot model=claude-opus-5.5 copilot=1.0.88 |
hanthor
left a comment
There was a problem hiding this comment.
Looks good. It needs a rebase onto main, and its base is older than the other open PRs'. Please confirm that avahi-gobject-devel, gtk-doc and xmlto are available in the build root.
Import phodav to supply spice-webdavd, lock its upstream archive, and register the source in the factory inventory. Run the upstream Meson tests and preserve configuration failures. Record recipe accounting separately from guest runtime acceptance. Assisted-by: GPT-6 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Hive-Run: projectbluefin#490 Hive-Plan: issue-490-implementation-steps-2-5 Hive-Spec: issue-490#exact-scope-spice-webdavd Signed-off-by: thatguywob <thatguywob@users.noreply.github.com>
|
Addressed review feedback: ensured all commits carry a Signed-off-by trailer and refreshed head to 🐝 Hive Agent: 🐝 Hive Agent: |
|
@thatguywob this PR is back in your court ( Review asks:
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 (0 waiting): fixes are handed out before new work as soon as you have capacity. |
|
hive-base-moved: merged base commit b650d86 🐝 Hive Agent: |
|
@thatguywob this PR is back in your court ( Review asks:
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. Your relay session received this as a |
Signed-off-by: thatguywob <thatguywob@users.noreply.github.com>
744393f to
13dede2
Compare
|
@thatguywob this PR is back in your court ( Review asks:
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. Your relay session received this as a |
|
@thatguywob the hive has adopted this PR ( |
Refresh architecture and backlog state from the current inventory. Add one inventory regeneration command, repair invalid imported YAML whitespace, and preserve workflow output values with grouped redirection. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
|
stage 080-awaiting-merge-admission → 074-adopted-by-hive: contributor-lane:changes-requested |
Regenerate combined 680-recipe architecture and backlog state after projectbluefin#847 landed on main, and remove the prohibited historical DELTA session artifact carrying committed conflict markers. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
Preserve phodav recipe and source lock alongside uvg266, regenerate combined 681-recipe inventory, and remove the prohibited historical session artifact. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
Merge current main (a590886) preserving all PR recipes and source locks, and regenerate combined inventory with just regenerate-inventory. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
|
Merged current |
|
@thatguywob this PR has a new head without confirmed hive push evidence, so the hive has handed this PR back to you ( |
|
@thatguywob 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. |
Drop the style-only grouped GITHUB_OUTPUT edit carried in from the interim baseline branch; main intentionally kept the original workflow so recipe PRs do not touch pipeline paths. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
|
Follow-up |
|
@thatguywob this PR is back in your court ( Review asks:
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. |
|
@thatguywob the hive has adopted this PR ( |
There was a problem hiding this comment.
HUMAN DECISION NEEDED — @thatguywob: the 27-file update pushed after the last review (head e3ae451) could not be read by the automated reviewer; a person must review that delta before merge.
Summary: This pull request adds the virtual machine file-sharing service. An earlier version was reviewed and approved automatically. Since then the author pushed a further update touching 27 files (about 1,200 changed lines), and the automated reviewer hit GitHub's hourly request limit before it could download that update, so none of the new changes have been checked. This note records that gap so the update is not merged on the strength of the earlier approval.
What I need from you: A maintainer should review the delta since the previously approved head by hand, or re-queue this PR once the request limit resets.
Scope of this note
- Previously reviewed head:
13dede2e7a0837066cd95c930c3681048fde5362(approve on all perspectives, no findings). - Current head:
e3ae451d5fb37e2c79b30b8aa26957ff22c6c383. - Delta per queue metadata: 27 files / 1171 lines, including
.packit.yaml,Justfile,config/factory-build-backlog.toml,config/upstream-sources.json,docs/architecture.md,docs/contributing.md, newpackages/libpaper/*(spec, two patches,localepaper.c,paperconf.1),packages/rygel/rygel.spec,packages/uvg266/*(spec, patch),reports/factory-build-backlog.json. - Not read: the diff for that delta. GitHub returned HTTP 429 (hourly cap) and no local copy of the head commit exists; the local evidence bundle for this head carries only CI status (all checks success) and no file content.
- Consequence: correctness, security, intent-alignment, style and docs-currency are all unassessed for the delta. In particular the new C source, patches and spec files under
packages/libpaper/andpackages/uvg266/have not been inspected, and the delta's file list (three unrelated packages plus factory backlog config) looks wider than the PR title "add the virtual machine file-sharing service" — a human should confirm whether that scope is intended or a stale branch merge.
No findings are asserted; nothing above cites code, because none could be read.
Confidence: 3/5 (needs attention) — 5 perspectives requires a human decision
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
Merge current main and regenerate derived inventory with just regenerate-inventory. Assisted-by: GPT-6.1 via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
|
@thatguywob this PR has a new head without confirmed hive push evidence, so the hive has handed this PR back to you ( |
|
@thatguywob this PR is back in your court ( Review asks:
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. |
Summary: This adds build instructions for a service that lets a virtual machine access a folder shared by its host. It supplies one missing part of the requested system tools, with verified source downloads and automated upstream tests. A package build and a file-transfer test are still needed before users can rely on it.
What I need from you: Review the recipe and run the trusted package build before merging.
Changes
phodav3.0 from Fedora dist-git commit6f5ddf3d09f950b1b443da3e7af60f840d660854(treebcc4ae728eb878c3f63e7462a4bf7a622b570ab7). It producesspice-webdavd,libphodav,libphodav-devel, andchezdav; no separatespice-webdavdsource recipe exists.%checkwith%meson_testto run upstream's virtual-directory test. Remove the imported configure command's|| catfallback so configuration errors fail the build.spice-webdavdinto the recipe-import[resolved]record and regenerate the backlog report, preserving the original audited names. This does not establish publication, dependency closure, or guest functionality.Refs #490 — partial source-family work; the other unresolved outputs and required build, publication, dependency-resolution, and runtime evidence remain. Consumer tracking: projectbluefin/utah#392 and projectbluefin/utah#263.
Validation
Review follow-up: Source0 now uses HTTPS with the original SHA-512 lock unchanged. The branch is based on current
main(235c5e88afd44348e835b45d70af5588edf8384a).Verified build dependency publication on 2026-10-31 against repositories enabled by
tools/mock_config.py: Hummingbird x86_64 publishesavahi-gobject-devel-0:0.9~rc4-0.1.2.hum1.x86_64; Fedora 44 Everything publishes gtk-doc-1.35.1-3.fc44.x86_64.rpm and xmlto-0.0.29-5.fc44.x86_64.rpm. Hummingbird repomd SHA-256:c84ca06d9d86dc7fa9b36d4e43dc0668a409c611ff6b206efc64f62fdd8885f8; compressed primary SHA-256 (verified against repomd):5a3be07b202357809eacefc5decd7e4d2e9eb24de081aefb46b8208495e33b39. These are publication checks, not a successful Mock lock or build; those and representative guest transfer remain required on parity(dev-tools): resolve factory outputs qemu-guest-agent through sssd-nfs-idmap #490.Downloaded
https://download.gnome.org/sources/phodav/3.0/phodav-3.0.tar.xz; its SHA-512 matches the importedsourcesmanifest and new lock:a4720083fd581023ea78cc09e6df4b5d2f90894f916004cf28c7fa557da61d8c7a5539fa6b50647efcd9f771afd088889fdc40579a8e860c8ea20540b25eb26a.just check: passed, including the backlog snapshot check and full unit suite.just test: passed (1,406 passed, 2 skipped).git diff --cached --check: passed.pre-commit run --all-files: unavailable (pre-commit: command not found). No checks were weakened.Attempted
gh workflow run canary.yml --repo thatguywob/utah-packages --ref hive/490-spice-webdavd -f 'lane_packages=["phodav"]' -f lane_backend=hermetic. Hive's command gate denied workflow dispatch (surface_allowlist); no build started and no alternate route was attempted.No built NEVRA, new factory digest, installation transaction, or successful guest file transfer is claimed. The upstream tests still need execution in the GitHub-hosted, digest-pinned build environment. Then resolve the built RPMs against only Hummingbird plus the factory and test file transfer in a SPICE guest with a WebDAV virtio channel and host-shared directory. A standalone
chezdavtest would not validate that channel.Batch disposition and publication snapshot
Inspected the published factory catalog on 2026-10-28. It was generated at
2026-10-28T03:33:30+00:00from factory imagesha256:218c483f4c5258b865244dc55203f6614e09519f05eb213ddb36bc596024ec23, revisionbb16990704ba4305844cc34a5c886961954adbc5(verified onupstream/main), publication run 40569848351. Catalog entries describe publication, not installability.Also fetched Hummingbird x86_64 repomd.xml, SHA-256
1e65e978ef506463967aa33deb54604167c44834d27d3da384bd1f290ed58f85, and its compressed primary metadata, SHA-2562f3ec264cd84e3113468065a391ac18a6697ea8138c9e66f97beae02b8eeb698. Of the batch's names, onlyrsyncappears there. Hummingbird also publishes the recipe's explicit runtime dependencyavahi; the factory publisheslibsoup33.7.3-1.hum1.bfin. Actual generated library requirements and solver closure remain unverified until the RPMs build.Utah's triage at
ccbc635713674629a31518050ac4e78915d8f2f7marksspice-webdavdplanned under guest integration; this PR does not change consumer policy.spice-webdavdphodav: absent from both snapshots; recipe and lock added here. Build, publication, dependency closure and guest test remain.rsyncrsync: already recorded in the catalog and published as 3.5.1-1.hum1.bfin; Hummingbird also carries 3.5.1-1.hum1. No change here.smartmontools,smartmontools-selinuxsmartmontools: already recorded and published as epoch 1, 7.5-11.hum1.bfin. No change here.sshpasssshpass: recipe already present and output published as 1.09-13.hum1.bfin. Runtime acceptance remains.spice-vdagentspice-vdagent: recipe already present and output published as 0.23.0-3.hum1.bfin. Guest runtime acceptance remains.qemu-guest-agent,qemu-user-static-aarch64rclone,realmd,resticrpm-ostree,rpm-ostree-libsrpm-ostreefamily: absent from both snapshots and remain in the backlog. No exclusion or consumer-policy decision here.sssd-client,sssd-common,sssd-kcm,sssd-krb5-common,sssd-nfs-idmapsssdfamily: absent from both snapshots and remain in the backlog. No identity-policy change here.— hive: backend=codex