Strategic Merge Patch Silently Converts Overriding Value Type to Match Original Type of Value Getting Overriden

[Magnitus-]

Files that can reproduce the issue

patch.yml:

apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
  name: letsencrypt
spec:
  privateKey:
    algorithm: RSA
    encoding: PKCS1
    size: 4096

source.yml:

apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
  name: letsencrypt
spec:
  privateKey:
    algorithm: __UNDEFINED__
    encoding: __UNDEFINED__
    size: __UNDEFINED__

kustomization.yaml:

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization

resources:
  - source.yml

patchesStrategicMerge:
  - patch.yml

In the directory with those files, run: kustomize build .

Expected output

Either:

• Output a type mismatch error (if you want to enforce overriding values type matching the type of the value getting overriden)
• Or output:

apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
  name: letsencrypt
spec:
  privateKey:
    algorithm: RSA
    encoding: PKCS1
    size: 4096

Actual output

apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
  name: letsencrypt
spec:
  privateKey:
    algorithm: RSA
    encoding: PKCS1
    size: "4096"

Kustomize version

3.8.4 and 3.8.7.

The problem was not observable in 3.6.x, it got introduced later.

Platform

Linux

[Shell32-Natsu]

I suspect this is caused by a lot of conversions among Json/Yaml/Go struct. The type information is lost. Hopefully we can avoid this after #2506 and remove all these intermediate conversions.

[fejta-bot]

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-contributor-experience at kubernetes/community.
/lifecycle stale

[fejta-bot]

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-contributor-experience at kubernetes/community.
/lifecycle rotten

[fejta-bot]

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-contributor-experience at kubernetes/community.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

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-contributor-experience at kubernetes/community.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.