Set list w/ atomic map types is not yet supported

[yue9944882]

structured-merge-diff/typed/helpers.go

Lines 211 to 213 in 1791273

	case child.IsMap():
	// TODO: atomic maps should be acceptable.
	return pe, errors.New("associative list without keys has an element that's a map type")

i was hitting above line when using an atomic map as the element of a set-list. is there a plan for supporting such schema? thanks! @apelisse @jennybuckley

[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-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[apelisse]

@jpbetz Good question, we should definitely add a test for that

[jpbetz]

@yue9944882 What was your exact case?

[yue9944882]

@apelisse @jpbetz thanks for reviving this thread. for my case, an example will be kubefed's model deifinition:

https://github.com/kubernetes-sigs/kubefed/blob/fd10623848151303a8e9e0d60a0dcace49cf9c43/example/sample1/federatedconfigmap.yaml#L16-L19

the model has an array of json-patches which is a map of op, path, value

[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-testing, kubernetes/test-infra and/or fejta.
/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-testing, kubernetes/test-infra and/or fejta.
/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-testing, kubernetes/test-infra and/or fejta.
/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.

[yue9944882]

/reopen
/lifecycle frozen

[k8s-ci-robot]

@yue9944882: Reopened this issue.

In response to this:

/reopen
/lifecycle frozen

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.

[apelisse]

Thanks @yue9944882, I think @jpbetz solved this one.

Feel free to re-open if I'm wrong.
/close

[k8s-ci-robot]

@apelisse: Closing this issue.

In response to this:

Thanks @yue9944882, I think @jpbetz solved this one.

Feel free to re-open if I'm wrong.
/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.

[apelisse]

/reopen

This is clearly not implemented yet :-)

[k8s-ci-robot]

@apelisse: Reopened this issue.

In response to this:

/reopen

This is clearly not implemented yet :-)

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.

[apelisse]

I don't think it's specifically super hard to implement. I wouldn't be completely opposed to forbidding it also (I'm not completely convinced we have a good use-case for that). What does the documentation say about this (for CRDs)?

[apelisse]

https://v1-18.docs.kubernetes.io/docs/reference/using-api/api-concepts/#merge-strategy

[apelisse]

i was hitting above line when using an atomic map as the element of a set-list. is there a plan for supporting such schema?

I'd love to understand the use-case to know if it's really useful, I have concerns because:

1. The current documentation says that's its not possible,
2. We have to save the entire atomic element in the managedFields if they are part of a set (to know who owns which one). If people start using it by accident, they may break things in unexpected ways.