Skip to content

fleet: busbar pin 7356bbfb6b (predev, the #496 merge) - #23

Merged
MattJackson merged 3 commits into
devfrom
p4-predev
Oct 7, 2026
Merged

MattJackson merged 3 commits into
devfrom
p4-predev

Conversation

@MattJackson

Copy link
Copy Markdown
Contributor

Repins to busbar predev 7356bbfb6b (the #496 merge).

  • scripts/fleet/repin.sh. Cargo.lock re-derived from busbar's lock at the pin.
  • Cargo.toml, deny.toml and release.yml are the busbar-release plugin render at the pin (busbar-release 6155dc1).
  • xtask plugin-gates parity and declares at the pin: 0 findings. cargo deny check: clean.
  • The rendered conformance host adapter now fills DoorFacts.status_rows from the transport facts. The field is new in busbar-core-connector at the pin, and busbar's root fills it the same way (root/doors.rs). The busbar-release template needs the same line, or the next sync reverts it.
  • Local proof: fmt, clippy -D warnings, and every test that runs without a live Postgres (the live-only cases and the suite's three live arms are CI's).

…Cargo.lock re-derived from busbar's lock at the pin; Cargo.toml, deny.toml and release.yml the busbar-release render at the pin

xtask plugin-gates parity and declares at the pin: 0 findings; cargo deny: clean.
…ar-core-connector at the pin), as the root's doors.rs fills them

At the pin busbar-core-connector's DoorFacts has status_rows (each claim's status code ranges); the
rendered host adapter is filled from the transport facts the loader reads, as busbar's root does.
The busbar-release template needs the same line.
… keypair the built busbar embeds), as the host grants this store's operator-infrastructure tcp need to a first-party plugin only

At the pin busbar refuses a non-first-party plugin whose Statement declares a need in a first-party
egress class (sign::egress_grant, BUSBAR-6002), so the unsigned tarball the two boot tests packed
was no longer loaded. The busbar they build embeds a fixed test-only public key
(BUSBAR_RELEASE_PUBKEY, compile time, as busbar's signing gate does) and the tarball is signed with
its private half at the built busbar's own version (the first-party anti-downgrade floor). Proven
locally against a live Postgres: the workspace's tests all pass.
@MattJackson
MattJackson merged commit 5e13886 into dev Oct 7, 2026
5 checks passed
@MattJackson
MattJackson deleted the p4-predev branch October 7, 2026 08:58
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