Skip to content

The decision records, the broker design and the supply-chain notes are shorter, with their claims kept - #247

Merged
DevomB merged 1 commit into
mainfrom
comments-docs
Oct 7, 2026
Merged

DevomB merged 1 commit into
mainfrom
comments-docs

Conversation

@DevomB

@DevomB DevomB commented Oct 7, 2026

Copy link
Copy Markdown
Owner

Part of the cleanup-held redo, after #241-#244 and #246. Prose only, in docs/decisions.md, docs/design/broker.md and docs/supply-chain.md. No heading changes, so every link into these files still resolves.

How it was done. Cleanup-held's change was merged onto main with git merge-file. Where main had changed a passage since, main's text stands. Each remaining cut was read against main, and against the code wherever it states a fact.

Kept where the cut corrects main (supply-chain.md):

  • Intel's microcode comes from Intel's own archive, which is unsigned. Only device firmware and AMD's microcode come from linux-firmware (build/config/versions.env: V_INTEL_MICROCODE, "AMD's is in linux-firmware"). The firmware bullet now says so and links to the section on intel-microcode.
  • cmake-bin is held to Kitware's signed SHA-256 list, like the source tarball (the sums:cmake-${V_CMAKE}-SHA-256.txt.asc row in tools/fetch-sources.sh). Main's text said only that it was checked once against a published list.
  • The proxy rewrites the title and then the app_id, "from which the compositor draws the zone's border": the border comes from the app_id (dwl-zone-borders.py), and the sentence now ends on it.

Merged by hand: two passages take the cut's sentences and paragraph break but keep main's claims. These are broker.md's mutation-test bullet and supply-chain.md's signature-gate paragraph (key-provenance.tsv, and a note left untried when its signature could not be fetched).

Dropped from the cut, because main moved past it:

  • decisions.md's opening "Every decision here is accepted": main now holds proposed records.
  • ADR-011's reason for the cookies, "a machine whose firmware has no SMT to turn off". Main's reason is that plain nosmt lets root turn SMT back on through sysfs.
  • ADR-012's "2.7 GB root image": main says about 1.7 GB.
  • broker.md's "Coverage-guided fuzzing needs nightly Rust and belongs in a scheduled job": that job exists (.github/workflows/fuzz.yml).

How the run proves it. Nothing builds from these files, and no script reads their text. CI's checks run as usual.

…e shorter, with their claims kept: cleanup-held's cut of docs/decisions.md, docs/design/broker.md and docs/supply-chain.md, redone on main by hand

Where main changed a passage after the cut, main's text stands: the records
include proposed ones, nosmt's cookies stay because root can turn SMT back on,
the root image is about 1.7 GB, fuzz.yml exists, and the signature gate reads
key-provenance.tsv. supply-chain.md carries two corrections the cut made:
Intel's microcode comes from Intel's own unsigned archive, not linux-firmware,
and cmake-bin is held to Kitware's signed SHA-256 list like the source tarball.
@DevomB
DevomB merged commit 2122cc9 into main Oct 7, 2026
11 checks passed
@DevomB
DevomB deleted the comments-docs branch October 7, 2026 20:19
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.

1 participant