Generator only works if k8s.io/apimachinery is in the $GOPATH about code-generator HOT 15 CLOSED

I worked out the issue - this doesn't work unless k8s.io/apimachinery is in the GOPATH, if it is just vendored, then deepcopy can't find it. At the very least, this needs to be documented somewhere, but it would be great if the vendored version worked as well.

Changed title to represent the actual issue.

from code-generator.

@sttts you can copy the above section into a Dockerfile and do a docker build . and see the failure in action 😄

from code-generator.

The cause for the issue described above is that the generators are not called from the sample-apiserver directory. Only then gengo (the Golang parsing library below) can pick up the vendor/ directory properly. The following Dockerfile passes the deepcopy-gen call, but fails at client-gen (compare kubernetes/kubernetes#57143 for a fix).

FROM golang:1.9.2

ENV PATH=$PATH:/go/bin
ENV GOPATH=/go
RUN go get -u k8s.io/code-generator/...
RUN go get -d k8s.io/sample-controller
RUN ls -l /go/src/k8s.io
RUN mkdir -p /go/src/k8s.io/kubernetes/hack/boilerplate && touch /go/src/k8s.io/kubernetes/hack/boilerplate/boilerplate.go.txt
WORKDIR /go/src/k8s.io/sample-controller
RUN CODEGEN_PKG=../../k8s.io/code-generator bash -xe hack/update-codegen.sh

from code-generator.

@markmandel please consider reopening this issue. I see no evidence that it has been solved, and users are still running into issues. Thank you.

from code-generator.

+1 for fix or documentation.

I encounter the same issue and only find the solution mentioned before by Google.

from code-generator.

@markmandel @jinghzhu do you have a small example repo where this can be seen?

from code-generator.

PR welcome to improve our codegen shell scripts to cope with out-of-dir invocations.

from code-generator.

Also note that the package is called k8s.io/sample-controller, not github.com/kubernetes/sample-controller.

from code-generator.

I am facing the same issue

from code-generator.

@rahulkrishnanfs compare my comment #21 (comment). You have to call the generator script with a relative CODEGEN_PKG. Does it help?

from code-generator.

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

from code-generator.

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten
/remove-lifecycle stale

from code-generator.

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

from code-generator.

/reopen

See kubernetes/kubernetes#79149 (comment)
Code generator users are still running into this issue.

Either the documentation should inform users to prepare their GOPATHS with the necessary dependencies or the code-generator be improved so it no longer has this requirement.

from code-generator.

@mattysweeps: You can't reopen an issue/PR unless you authored it or you are a collaborator.

from code-generator.