Drop k8s third_party deps now that HEAD is updated - #1480
Merged
Benjamin Elder (BenTheElder) merged 10 commits intoSep 5, 2026
Merged
Drop k8s third_party deps now that HEAD is updated#1480Benjamin Elder (BenTheElder) merged 10 commits into
Benjamin Elder (BenTheElder) merged 10 commits into
Conversation
The changes we need are released now (sort of - on HEAD anyway).
GOPROXY=direct go get \
k8s.io/apimachinery@master \
k8s.io/code-generator@master
This pins them to HEAD of master. Go is terrible here: The HEAD is not
actually tagged, so Go just uses the next "reachable" tag which is
v0.36.0-alpha. The datestamp is correct, though.
The changes we need are released now (sort of - on HEAD anyway).
GOPROXY=direct go mod edit -replace k8s.io/apimachinery=k8s.io/apimachinery@master GOPROXY=direct go mod tidy GOPROXY=direct go mod vendor GOPROXY=direct go mod tidy This approach (-replace) is needed because Go is horrible here. The master branch of k8s.io/apimachinery is not tagged, per se, but there is an OLDER tag which is "reachable" from HEAD. So go helpfully decides to use that (v0.36.0-alpha.2). If we just `go get ... @master` it works for that dep (pinned to the right date) but then it looks at transitive deps. Because the tag seems to be 0.36 (older), it recalculates all the OTHER dependencies and downgrades a whole tangle of things to versions that match 0.36, but we are ACTUALLY on 0.37+. This was the only approach that I (and Gemini) could find. Blech.
Collaborator
|
I think you need to regenerate the license folder / manually fix it, to drop the dead third_party licenses |
Tim Hockin (thockin)
force-pushed
the
dv_update_to_head
branch
from
September 4, 2026 17:00
1c7bc0f to
2214c0b
Compare
| k8s.io/apimachinery v0.37.0-rc.0 // indirect | ||
| k8s.io/code-generator v0.37.0-rc.0 // indirect | ||
| k8s.io/apimachinery v0.36.0-alpha.2.0.20260904054801-5a60cf1c6b81 // indirect | ||
| k8s.io/code-generator v0.30.0-alpha.3.0.20260904060942-0cddac4383e4 // indirect |
Collaborator
There was a problem hiding this comment.
that's weird, we should have at least v0.37.0-alpha ?? I wonder if the generators are failing to sync. not super important at the moment but ...
we can check with Davanum Srinivas (@dims) later
Collaborator
Author
There was a problem hiding this comment.
The datestamps are right but we don't tag on main and git's notion of "reachable" is not human-oriented
Collaborator
There was a problem hiding this comment.
I see https://github.com/kubernetes/code-generator/tree/v0.37.0 for sure
| @@ -0,0 +1,15 @@ | |||
| ISC License | |||
Collaborator
There was a problem hiding this comment.
we need to start enforcing a license allowlist at some point
I double checked, ISC is on the list (also I assume k8s is already using it anyhow, but k8s also has some exceptions I think)
https://github.com/cncf/foundation/blob/main/policies-guidance/allowed-third-party-license-policy.md
Collaborator
Author
|
green |
Benjamin Elder (BenTheElder)
approved these changes
Sep 5, 2026
Benjamin Elder (BenTheElder)
merged commit Sep 5, 2026
2fcfa64
into
agent-substrate:main
11 of 13 checks passed
Max Smythe (maxsmythe)
pushed a commit
that referenced
this pull request
Sep 5, 2026
`main` is red since #1480 merged: its regenerated `cmd/ateapi/internal/controlapi/zz_generated.validation.go` was produced by a stale generator build and references `deepEqualImpl_`, which doesn't exist — `cmd/ateapi` (and everything importing `controlapi`) fails to compile, and every open PR inherits the failure through CI's merge-with-main (see #1512's `run-tests` for an example, and `main`'s own post-merge `pr-workflow` runs for #1480/#1499). Fix is a pure regeneration: `hack/update/codegen.sh` on current `main` changes exactly one line, `deepEqualImpl_` → `ateDeepEqual` (the helper the current generator actually emits). After it, `go build ./cmd/ateapi/...` succeeds and `go test ./cmd/ateapi/internal/controlapi/` passes. No hand-written changes — the diff is generator output only. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The changes we need are live on HEAD of the k8s published repos. It was a bit of a journey, since Go fought me all the way.
Drop our third_party fork of k8s deps in tools
Bump k8s.io/streaming to v0.37.0 (not rc) in tools
Bump codegen deps to HEAD in tools
Drop our third_party fork of k8s deps in root
Bump k8s deps to v0.37.0 (not rc) in root
Bump apimachinery dep in root to HEAD in root
Run updated codegens
Use DV's new maxBytes capability for
[]byte