Skip to content

chore(companions): new versions carrying a codefly with the delivery verbs - #875

Merged
antoinetoussaint-byte merged 1 commit into
mainfrom
publish-codefly-0.0.7
Oct 29, 2026
Merged

antoinetoussaint-byte merged 1 commit into
mainfrom
publish-codefly-0.0.7

Conversation

@antoinetoussaint-byte

Copy link
Copy Markdown
Contributor

Why

Every companion bakes the codefly CLI. The CLI gains codefly delivery prepare|apply (codefly-dev/cli branch handoff/wiki-builder-install-20261029): a cell's delivery Job runs apply from the codefly companion to send a module's signed pair — the Core carrier and the authority declaration signed against it — to the host's codefly.dev/composition interface. That is what lets a GitOps rollout deliver a module's roles; the curl runner's REST route carries none, which is why Blue's accounts refuses to compose wiki ("0 roles and none installer") and the install review answers module "wiki": not found.

The published codefly:0.0.6 carries a codefly built before the verb existed, so the Job cannot run it. Publishing skips a tag already in the registry, and --force would swap the image under every consumer that has resolved 0.0.6, so this is a version bump — and because every companion carries the CLI, every one of them bumps (the runbook's own rule).

What

companion from to
codefly 0.0.6 0.0.7
execution 0.0.2 0.0.3
go 0.0.13 0.0.14
node 0.0.15 0.0.16
proto 0.0.23 0.0.24
python 0.0.8 0.0.9

The four dependents' CODEFLY_BASE_IMAGE defaults follow the base to 0.0.7 (buildspec_test enforces it). No other Dockerfile input changes.

Classification

chore, at the place that owns companion versions. Nothing reused or built; the publish itself is the existing CLI Companions workflow (workflow_dispatch, core_ref = this commit), run from the CLI ref carrying the verbs, as docs/runbooks/publish-companions.md describes.

Testing

  • go test ./companions/... ./internal/ciguard/... pass.
  • go run ./internal/companionbump ghhttps/main: every companion whose image changed carries a new version.
  • Not verified here: the images themselves; the PR's companions-build builds every image from this branch, and the publish records the digests the consumers then pin.

Refs obin-ai/infra-base#1509 (Blue delivery).

🤖 Generated with Claude Code

…he delivery verbs

Every companion bakes the codefly CLI, so a CLI change that an image must run
needs a new tag for every one of them: publishing skips a tag the registry
already has, and overwriting one would swap the image under every consumer
that has resolved it. The CLI gains `codefly delivery prepare|apply`, which a
cell's delivery Job runs from the `codefly` companion to send a module's signed
pair (Core carrier plus the authority declaration) to the host's composition
interface; the 0.0.6 image carries a codefly without that verb, so the Job
cannot run it.

codefly 0.0.6 -> 0.0.7, execution 0.0.2 -> 0.0.3, go 0.0.13 -> 0.0.14,
node 0.0.15 -> 0.0.16, proto 0.0.23 -> 0.0.24, python 0.0.8 -> 0.0.9. The four
dependents' CODEFLY_BASE_IMAGE defaults track the base, as buildspec_test
requires. No Dockerfile input changes otherwise.

Testing: go test ./companions/... and ./internal/ciguard/... pass;
`go run ./internal/companionbump ghhttps/main` reports every changed companion
carries a new version. Publishing is the CLI's Companions workflow against this
commit, dispatched from the CLI ref that carries the verbs.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
antoinetoussaint-byte added a commit to codefly-dev/cli that referenced this pull request Oct 29, 2026
…e first that carries the verb

pairedDeliveryImage pinned companion 0.0.6, whose static codefly predates
`delivery apply`, so a paired Job rendered against it failed on an unknown
flag. Every companion bakes the CLI, so core bumped them all (codefly-dev/core#875)
and the Companions workflow published 0.0.7 from this branch; its index digest
is sha256:1cbd3826…, measured on linux/amd64 and linux/arm64 to answer
`delivery apply --help`.

A moved pin changes every cell's delivery declaration, so the cell is
re-rendered with it; the docs say that instead of "not yet runnable".

Testing: GOWORK=off go test ./pkg/gitops/ ok (222s).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
@antoinetoussaint-byte
antoinetoussaint-byte enabled auto-merge (squash) October 29, 2026 17:39
@antoinetoussaint-byte
antoinetoussaint-byte merged commit 15e7e12 into main Oct 29, 2026
17 checks passed
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