Prepare clock package to be compatible replacement of apimachinery/util/clock

[MadhavJivrajani]

What type of PR is this?
/kind cleanup
What this PR does / why we need it:
Commit 1:
The reason for doing this is that utils/clock is supposed
to be a full replacement for k8s.io/apimachinery/pkg/util/clock.

Currently, WithDelayedExecution extends Clock - in order to deprecate
apimachinery/clock, WithDelayedExecution would be missing the NewTicker
method if gradual deprecation via type aliasing was to be done.

By extending WithTicker instead of Clock, we solve this issue.

Commit 2:
Eventhough the implementation is just a panic,
re-implementing the NewTicker method helps
in migration of the apimachinery/util/clock pkg
to the utils/clock pkg in terms of having identical
clients for IntervalClock in both packages as
well as the IntervalClock implementing the
WithTickerAndDelayedExcution interface.

Which issue(s) this PR fixes:
Related to kubernetes/kubernetes#94738

Special notes for your reviewer:
The next step after this would be to implement the Tick method in apimachinery clock, type aliasing the interfaces and marking the package as deprecated.

Release note:

NONE

/assign @liggitt @dims
/cc @wojtek-t

[liggitt]

this has the same problem with making incompatible changes to existing interfaces as #221 (comment)

[MadhavJivrajani]

The API diff will fail here. Do we introduce a new interface to further encapsulate things?
Considering that this change was done only a few days back, is it possible to somehow get this in or would there be a better way to go about it?

[wojtek-t]

this has the same problem with making incompatible changes to existing interfaces as #221 (comment)

@liggitt - This is something we may want to consider, as this type was introduced couple days ago.
I don't have strong opinion though - just pointing out.

[liggitt]

@liggitt - This is something we may want to consider, as this type was introduced couple days ago.

merged in #221 ~2 weeks ago, right?

[MadhavJivrajani]

Yep :)

[MadhavJivrajani]

Hey @liggitt - any updates on this?

[liggitt]

once an interface is merged, available, and picked up as a dependency in other places, we should not break the interface

https://grep.app/search?q=WithDelayedExecution shows the existing interface has already been picked up by some repos other than kubernetes/kubernetes

[MadhavJivrajani]

Ah I see - thanks for pointing towards that, I'll revert the change and introduce another interface.
Sorry about the back and forth on this

[MadhavJivrajani]

xposting: kubernetes/kubernetes#94738 (comment)

[wojtek-t]

/lgtm

@liggitt @dims - can you take a look?

[MadhavJivrajani]

/hold
(would like to double check once again post approval)

[dims]

/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: dims, MadhavJivrajani, wojtek-t

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:

• OWNERS [dims]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[liggitt]

can you open a WIP PR against kubernetes/kubernetes that bumps k8s.io/utils to this commit and ensures we can completely switch over to this package and get green CI on that PR?

you can do that in kubernetes/kubernetes like this:

hack/pin-dependency.sh k8s.io/utils=github.com/MadhavJivrajani/utils rectify-interface-extension
hack/update-vendor.sh

[MadhavJivrajani]

@liggitt @wojtek-t fyi - kubernetes/kubernetes#106850 :)

[wojtek-t]

That LGTM

[MadhavJivrajani]

Thank you, both!
Removing hold now seeing that the CI in the test PR is green :)
/hold cancel