update exit code to 0 if patch not needed #66725
Conversation
k8s-ci-robot
added
release-note-none
k8s-ci-robot
added
release-note
|
/sig cli |
k8s-ci-robot
added
the
sig/cli
| @@ -245,14 +245,8 @@ func (o *PatchOptions) RunPatch() error { | |||
| if err != nil { | |||
| return err | |||
| } | |||
| printer.PrintObj(info.Object, o.Out) | |||
There was a problem hiding this comment.
Can you please add a tests verifying:
- that exit code is 0, when the patch was empty
- that exit code is non-zero, when the patch command actually failed
There was a problem hiding this comment.
@soltysh Added a test. Although I realize after adding it, that only semantic or type errors will result in the error being printed out and an exit code of 1. If someone were to, for example, update an immutable field on a pod, or set a value for a non-existent spec field on a pod (like replicas), the server would just return the old object with no error, and the command would exit with the message "not patched" (as before), but with exit code 0 (since the old and new object were the same)
k8s-ci-robot
added
size/S
k8s-ci-robot
added
the
approved
juanvallejo
force-pushed
the
jvallejo/update-patch-return-code-logic
branch
2 times, most recently
from
91c0f0d to
578ab90
Compare
|
/test pull-kubernetes-integration |
juanvallejo
force-pushed
the
jvallejo/update-patch-return-code-logic
branch
from
578ab90 to
ad11a1b
Compare
k8s-ci-robot
added
the
lgtm
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: juanvallejo, soltysh The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/test all [submit-queue is verifying that this PR is safe to merge] |
|
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here. |
Release note:
The specific logic in the
patchcommand that exited with code 1, was only doing so when there was no diff between an existing object and its patched counterpart. (In case of errors, we just return those, which eventually ends up exiting with code 1 anyway). This patch removes this block, as we should not be treating patch no-ops as errors.Fixes #58212
cc @soltysh