Skip to content

fix: install/arch leftovers (#11023 #11025) - #12932

Open
kvnloo wants to merge 16 commits into
omacom:quattrofrom
kvnloo:salvage/t1-install-arch-recut-20260922
Open

kvnloo wants to merge 16 commits into
omacom:quattrofrom
kvnloo:salvage/t1-install-arch-recut-20260922

Conversation

@kvnloo

@kvnloo kvnloo commented Sep 22, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Re-cut of install/arch leftovers onto current quattro after #11501 closed unmerged.

Cherry-picks unique hunks from:

Out of scope: #11024 (already closed). Shell/tray leftovers are a separate PR.

Why still needed

On quattro @ d3cfd53 these paths were still absent:

  • bin/omarchy-root-btrfs-extra-luks
  • bin/omarchy-hw-x86-64

default/omarchy/omarchy-menu.jsonc drifted — conflict resolved here; Install rows stay gated by arch/prereq helpers only.

Test plan

  • ./test/shell — root-btrfs-extra-luks-test.sh (12/12), hw-x86-64-test.sh (3/3), menu-test.sh (incl. x86_64-only Install gates)
  • aarch64: x86_64-only Install rows hidden
  • x86_64: those rows visible; still dim when already installed
  • multi-device LUKS Btrfs root: warn fires; single-device: silent

Linear: PER-1528. Source heads: close #11023 #11025 as salvaged after this lands.

@kvnloo kvnloo changed the title salvage(install-arch): leftover cluster (#11023 #11025) fix: install/arch leftovers (#11023 #11025) Sep 22, 2026
@llstrk

llstrk commented Sep 22, 2026

Copy link
Copy Markdown

No important actionable production defect was found in the tested scope at fb1ba61e.

Verified: The seven changed Install rows retain their installed-state dimming on x86_64 and are hidden for modeled aarch64/arm64 systems. Checks exercised the production menu model and generated guards with synthetic architecture and package-presence responses.

Modeled machine/state Base PR
x86_64, absent 7 selectable rows 7 selectable rows
x86_64, installed 7 dimmed rows 7 dimmed rows
aarch64/arm64, either state 7 visible rows 0 visible rows

Verified: With synthetic storage responses, a distinct extra LUKS-backed Btrfs member produces the warning; single-member, plain-extra-member and same-LUKS-container cases remain silent. Root aliases and subvolume suffixes were covered. The production update-log analyzer preserves the warning without failing:

Modeled root: LUKS A + LUKS B
  helper detects distinct encrypted member
    warning on stderr, exit 1
  analyzer's `|| true`
    warning retained, exit 0

Review information

Test scope: Targeted base/head menu and helper checks, update-log integration, and source inspection. No rendered desktop, native ARM installation, complete ARM package-availability audit, real multi-device Btrfs boot, or full repository suite. Storage detection used stubs; native enumeration permissions were not validated. The warning assumes the stock unlock setup and does not inspect a customized multi-device initramfs configuration.

Community review: Independent automated community review, unaffiliated with the Omarchy team, intended to help prepare PRs for their review.

Automated AI review: Astra initial inspection, independent Opus 5.5 High and Astra Xhigh technical reviews, followed by Astra adjudication of the reports and targeted evidence.

@kvnloo

kvnloo commented Sep 22, 2026

Copy link
Copy Markdown
Contributor Author

Verified against current quattro (fail-then-pass on scoped shell tests). Ready for maintainer review.

Scoped: root-btrfs-extra-luks-test, hw-x86-64-test, menu-test.

@jandrusk

jandrusk commented Oct 8, 2026

Copy link
Copy Markdown

Thanks for re-cutting these. Two notes:

  • NIT, omarchy-root-btrfs-extra-luks: lsblk -nspo NAME,FSTYPE prints in tree form, so ancestor names come back with └─ prefixes whose depth depends on the stack. Two members on the same LUKS container at different depths (for example one directly on it and one through LVM) would then compare as different and warn falsely. List mode avoids that:

    done < <(lsblk -nslpo NAME,FSTYPE "$device" 2>/dev/null)

    The test stubs lsblk with flat output, so it can't catch this.

  • Scope: the extra-LUKS warning and the x86_64-only Install gating are unrelated changes. Splitting them would let each land on its own. Which rows count as x86_64-only (Zed, Cursor, Ollama, and so on) is also a product call that may change as omarchy-pkgs adds aarch64 builds.

@birkskyum

Copy link
Copy Markdown
Contributor

The Install gating here has gone stale on quattro. Ollama, LM Studio, Cursor, Zed and Symfony CLI all resolve on aarch64 now, and Edge, Dropbox and Spotify have been gated by omarchy-hw-x86 since #14431, so omarchy-hw-x86-64 and the menu hunks can go. #15753 hides the three Install rows that still can't work on ARM (GeForce NOW, Lutris, Battle.net). That leaves the Btrfs extra-LUKS warning as this PR's own change.

@greptile-apps

greptile-apps Bot commented Oct 18, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerView in GreptileConfidence Score: 5/5

[Medium impact] The PR appears safe to merge; no outstanding blocking issues were found.

Summary

Adds an advisory warning when root Btrfs spans separate LUKS-backed devices.

  • Updates warn when root Btrfs needs another LUKS unlock.

Reviews (3) · Last reviewed commit: "test(storage): exercise unencrypted Btrf..." · Reviewed by Greptile

Comment thread bin/omarchy-root-btrfs-extra-luks Resolved
Comment thread bin/omarchy-root-btrfs-extra-luks Resolved
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants