Skip to content

feat(packages): install what Bluefin inherits from Fedora's base and Hummingbird already has - #129

Merged
castrojo merged 1 commit into
mainfrom
claude/base-parity
Sep 18, 2026
Merged

castrojo merged 1 commit into
mainfrom
claude/base-parity

Conversation

@hanthor

@hanthor hanthor commented Sep 18, 2026

Copy link
Copy Markdown
Member

Why

The image-level parity report (#117) found 277 packages in Bluefin's image that Utah's lacks and that public-hummingbird-x86_64-rpms carries today. Bluefin never names them because Silverblue's base supplies them; Hummingbird's server-shaped bootable base leaves them out. This is the sweep that briefly rode on #105 and was lifted out at review (out of scope there, and its first version was red).

What changed

A new [parity] section in packages/utah.toml with 28 names: the NetworkManager connectivity check, alsa-ucm, langpacks-en and the core font set, and the shell tooling Bluefin users reach for (bash-completion, man-db, zip, python3-pip, ...). Dakota ships the same audio and font pieces.

Three names from the #105 version are gone because CI's real transaction rejected their closures, which a first-level Requires check had missed:

dropped why
NetworkManager-wwan needs ModemManager; the factory's build needs libmbim, libqmi and libqrtr-glib that nothing provides
NetworkManager-bluetooth needs libnm-wwan.so from the wwan plugin
poppler-utils Hummingbird's poppler needs poppler-data, which Hummingbird does not ship

They are recorded beside the list with the others that did not resolve (cups, vim-enhanced, the emoji and math fonts, fprintd, alsa-utils), each with its tracking issue (utah-packages#144 to #147). NetworkManager-wifi joins when the factory image from utah-packages#133 is pinned (#126).

unzip is listed and kept: it was already in [build], but configure-services.sh removed the build tooling after the extension build, so the image did not actually ship it. The removal now skips it and a test pins both halves.

install-packages.py adds the section to the install set, verify-rpm-contract.py asserts it on the built image, tests cover the install path, the verifier, and that [parity] duplicates no other section's claim. docs/skills/package-contract.md documents the section.

Verification

  • 68 unit tests pass locally; verify-rpm-contract.py --check reports 28 parity packages; bash -n clean.
  • CI's "Check package availability in Utah's repositories" step is the real gate: it resolves the transaction with dnf --assumeno against the pinned images.

Ordering with #117

31 of these names are in #117's parity-baseline.txt. Whichever lands second updates the other's file: if this merges first, #117 removes them from the baseline; if #117 merges first, this PR gets a follow-up commit doing the same.

🤖 Generated with Claude Code

https://claude.ai/code/session_01W6BYLUD6KBzF8rBFXKjP7M


Generated by Claude Code

…ady has

The image-level parity report (#117) found 277 packages in Bluefin's image
that Utah's lacks and that public-hummingbird-x86_64-rpms carries today:
Bluefin never names them because Silverblue's base supplies them, and
Hummingbird's server-shaped bootable base leaves them out. A new [parity]
section in packages/utah.toml adds the desktop-relevant ones -- the
NetworkManager connectivity check, ALSA use-case profiles, the English
language pack and core font set, and the shell tooling Bluefin users reach
for -- 28 names.

This is the sweep that briefly rode on #105 and was lifted out at review:
out of scope there, and its first version was red. Three names from that
version are gone because CI's real transaction rejected their closures,
which a first-level Requires check had missed: NetworkManager-wwan needs
ModemManager, whose factory build needs libmbim, libqmi and libqrtr-glib
that nothing provides; NetworkManager-bluetooth needs libnm-wwan.so from
the wwan plugin; poppler-utils needs poppler-data, which Hummingbird does
not ship. Each is recorded beside the list with the others that did not
resolve (cups, vim-enhanced, the emoji and math fonts, fprintd,
alsa-utils), each with its tracking issue.

unzip is listed and kept: it was already in [build], but
configure-services.sh removed the build tooling after the extension
build, so the image did not actually ship it. The removal now skips it
and a test pins both halves.

install-packages.py adds the section to the install set,
verify-rpm-contract asserts it on the built image, and the tests cover
the install path, the verifier, and that [parity] duplicates no other
section's claim. CI's package availability step is the gate on every
name here. Whichever of this and #117 lands second updates
parity-baseline.txt.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W6BYLUD6KBzF8rBFXKjP7M

@castrojo castrojo 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.

This is the split I asked for on #105, and it answers every point from that review. Thanks for turning it round that fast.

Confirmed against the original objections:

  • The three unresolvable names are gone, each documented with the specific dependency the pinned repositories do not carry: NetworkManager-wwan (ModemManager needs libmbim/libqmi/libqrtr — a factory gap), NetworkManager-bluetooth (falls with the wwan plugin's libnm-wwan.so), poppler-utils (Hummingbird's poppler wants poppler-data). That matches exactly what run 35363861546 reported. contract=SUCCESS here is the proof the remaining transaction actually resolves — and the header is right that a first-level Requires check would not have caught these.
  • unzip is fixed properly. It is in [parity], it is removed from the configure-services.sh teardown line, and the comment on that line explains why it is exempt. That closes the gap where the manifest claimed parity the image did not deliver.
  • Every other exclusion carries a tracking issue (utah-packages#144/#145/#146/#147), and NetworkManager-wifi is explicitly deferred to #133/#126 rather than silently dropped. That is the [unavailable]-discipline applied to a new section, which is the right instinct.
  • Scope is one logical change. 6 files, +123/-5.

Wiring checks out: [parity] is added to the install set in install-packages.py, classified and asserted in verify-rpm-contract.py (it is in expected, so a missing parity package fails the build rather than being skipped), the section is documented in both the utah.toml header and docs/skills/package-contract.md, and the count appears in the verifier's output line.

test_parity_section_reaches_the_install_set and test_parity_section_duplicates_nothing_but_the_build_tooling_it_keeps are both good — the second one is doing real work, asserting the unzip overlap is the only overlap and that the removal line actually omits it.

One nit, not blocking. test_verifier_asserts_the_parity_section greps the source of verify-rpm-contract.py:

self.assertIn('parity = section(overlay, "parity")', source)
self.assertIn("*parity,", source)

That asserts implementation text, not observable behaviour. It breaks on a harmless rename and would still pass if the classification were wrong. The behaviour it is reaching for is already covered by the other two tests; if you want it explicit, assert that a missing [parity] package makes the verifier exit nonzero. I would just delete it. Not worth another CI round on its own — fold it in if you touch this branch again.

Merge-order note: this closes parity-baseline entries that #117 records, so whichever of the two lands second needs its counts reconciled. #117 is currently open awaiting a second reviewer.

Approving on the strength of contract + main image + Trivy green; the three kernel-flavor images are still building, and I will merge once they finish rather than now.

@castrojo
castrojo merged commit 2216657 into main Sep 18, 2026
14 checks passed
@castrojo
castrojo deleted the claude/base-parity branch September 18, 2026 17:01
hanthor pushed a commit that referenced this pull request Sep 18, 2026
#116 (glibc-all-langpacks) and #129 (the [parity] section) landed after the
baseline was generated, closing 25 of its 969 names. By the file's own rule
they come out: 944 remain, 253 of them in Hummingbird's repository. The
pinned count in tests/test_image_parity.py follows. The header keeps the
Bluefin and Utah totals from the run that generated it and says so; the
next image build's report is where the refreshed Utah count comes from.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W6BYLUD6KBzF8rBFXKjP7M
hanthor pushed a commit that referenced this pull request Sep 18, 2026
main now carries the [parity] section (#129) beside this branch's
[hardware] section; the installer, the verifier and the tests keep both,
in the order gnome, hardware, parity, services.

The Bluetooth enablement comes out, as reviewed: #125 enables the same
unit and also adds it to contracts/bluefin-desktop.toml, which is what the
in-image desktop contract asserts, so it should own the fix rather than
have the duplicate hunk here race it. input-remapper stays.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W6BYLUD6KBzF8rBFXKjP7M
castrojo pushed a commit that referenced this pull request Sep 22, 2026
#116 (glibc-all-langpacks) and #129 (the [parity] section) landed after the
baseline was generated, closing 25 of its 969 names. By the file's own rule
they come out: 944 remain, 253 of them in Hummingbird's repository. The
pinned count in tests/test_image_parity.py follows. The header keeps the
Bluefin and Utah totals from the run that generated it and says so; the
next image build's report is where the refreshed Utah count comes from.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W6BYLUD6KBzF8rBFXKjP7M
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.

3 participants