FakeClock.Sleep calling FakeClock.Step may be unexpected

[SOF3]

What happened:

Consider the following code (playground):

package main

import (
	"time"

	"k8s.io/utils/clock"
	clocktesting "k8s.io/utils/clock/testing"
)

func main() {
	clk := clocktesting.NewFakeClock(time.Now())
	go func(clk clock.Clock) {
		clk.Sleep(time.Second * 2)
		println("after 2 fakeclock seconds")
	}(clk)

	time.Sleep(time.Millisecond) // used to emulate race condition
	clk.Sleep(time.Second)
	println("after 1 fakeclock second")
}

The code above prints

after 2 fakeclock seconds
after 1 fakeclock second

When clk is changed to clock.RealClock{} (playground), the println lines are in correct order (1 -> 2).

What you expected to happen:

Intuitively, one would expect clk.Sleep(duration) to be identical to <-clk.After(duration). However, (*FakeClock).Sleep actually calls clk.Step instead, which immediately modifies the clock.

Suggested fix:

It would be a huge BC break to many testing packages if (*FakeClock).Sleep is changed to be passive. Documentation improvement would suffice, although deprecating Sleep would be great since Sleep doesn't always do what users intend. (imo, clk.Step is usually only intended in the main control flow of the unit test rather than other goroutines, so calling clk.Sleep from non-testing code is often unintended).

Alternatively, consider adding a ref-counted clock, where Sleep and After have identical behavior and the clock is only stepped when all refs are sleeping.

Environment:
k8s.io/utils: fe8a2dd

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[SOF3]

/remove-lifecycle stale

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[SOF3]

/remove-lifecycle stale