CBG-4915: Fix flakiness of TestMultiActorLosingConflictUpdateRemovingAttachments

[RIT3shSapata]

CBG-4915

Describe your PR here...

• The test was flaky because the test would progress from the require.EventuallyWithT even though the condition is not completely met
• Refactored the code to use RestTester's WaitForVersion method instead

Pre-review checklist

• Removed debug logging (fmt.Print, log.Print, ...)
• Logging sensitive data? Make sure it's tagged (e.g. base.UD(docID), base.MD(dbName))
• Updated relevant information in the API specifications (such as endpoint descriptions, schemas, ...) in docs/api

Dependencies (if applicable)

• Link upstream PRs
• Update Go module dependencies when merged

Integration Tests

• GSI=true,xattrs=true https://jenkins.sgwdev.com/job/SyncGatewayIntegration/0000/

[Copilot]

Pull request overview

This PR fixes flakiness in the TestMultiActorLosingConflictUpdateRemovingAttachments test by replacing require.EventuallyWithT with RestTester.WaitForVersion() to properly wait for document replication.

Key Changes:

• Replaced manual polling loop with built-in WaitForVersion method for more reliable synchronization

[Copilot]

The error return value from GetDoc is being silently discarded. Consider handling the error or using require.NoError(t, err) to ensure the document retrieval succeeds before proceeding with the test.

Suggested change

	rtBVersion, _ := rtB.GetDoc(docID)
	rtBVersion, err := rtB.GetDoc(docID)
	require.NoError(t, err)

[torcolvin]

I think given this assertion, that rtBVersion and rtAVersion are the same.

I would perhaps assert that the version has CV and Revtree.

It is possible that this assertion will only work on CV, which will wait for the document to be copied (CV matches) but not imported.

I haven't looked into this, but we can discuss.

[torcolvin]

This needs to get changed too

[torcolvin]

We should move these to a defer. In the case of rosmar, the xdcr replications will stop when the buckets disappear but in CBS they won't, and so this could leak into future tests.

Maybe this is actually already handled by defer at the top.