docs: publishing a version here is only half of shipping it #247

Merged
binjovi-bot merged 1 commit from sean/consumer-pin into trunk 2026-09-09 13:20:17 +00:00 AGit
Owner

sean/binjovi copies tools out of these images pinned by tag and digest, so a new version published here reaches nothing until that pin is bumped there -- and nothing fails in the meantime. binjovi-buildkit-client v7 fixed the pull-request fetch on 2026-09-08 and the consumer stayed on v4 for a day, during which no pull request in this repository could build while trunk builds kept passing. This records the coupling on the side that publishes. It is also the canary for that fix: this pull request exists only at refs/pull, so building it at all is the proof. https://claude.ai/code/session_01KZoQin34jeyt6nDGqvJA76

sean/binjovi copies tools out of these images pinned by tag and digest, so a new version published here reaches nothing until that pin is bumped there -- and nothing fails in the meantime. binjovi-buildkit-client v7 fixed the pull-request fetch on 2026-09-08 and the consumer stayed on v4 for a day, during which no pull request in this repository could build while trunk builds kept passing. This records the coupling on the side that publishes. It is also the canary for that fix: this pull request exists only at refs/pull, so building it at all is the proof. https://claude.ai/code/session_01KZoQin34jeyt6nDGqvJA76
docs: publishing a version here is only half of shipping it
All checks were successful
binjovi/ci Binjovi completed the frozen plan
b716723c13
`sean/binjovi` copies tools out of several of these images at build time,
pinned by tag and digest. A new version published here reaches nothing until
that pin is bumped there, and nothing fails in the meantime: this repository
builds and publishes exactly as it should, and the consumer keeps running what
it was pinned to.

That gap has bitten once and it hid well. `binjovi-buildkit-client` v7 fixed
`binjovi-dockerfile-set` to fetch `+refs/pull/*/head`, without which the tool
cannot resolve a commit created by an agit push -- every pull request in this
fleet. v7 was built, published and signed on 2026-09-08 and `sean/binjovi`
stayed on v4 for a day. Trunk builds passed the whole time, because a plain
clone can see a trunk commit, so the lane looked healthy while no pull request
in THIS repository could build at all. Both open pull requests here went stale
waiting for a build that could never succeed.

The pin is fixed and `container_contract_test.exs` in `sean/binjovi` now holds
the minimum version with the reason. This records the coupling on the side that
publishes, because that is the side that knows a new version exists.

Claude-Session: https://claude.ai/code/session_01KZoQin34jeyt6nDGqvJA76
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
sean/builder-images!247
No description provided.