CBL-3774: Collection SG test, "Auto Purge Enabled - Revoke Access"

[callumbirks]

Port test. Several adjustments made to fit into this test suite.

[callumbirks]

Looks like I accidentally included a change in the comments at the top of the file, this change is also present though in the ongoing PR CBL-3794.

[cbl-bot]

Code Coverage Results:

Type	Percentage
functions	79.68
instantiations	33.52
lines	79.66
regions	75.32

[pasin]

PR Validation Fail. Please make sure that PR Validation are all passed before submitting for review.

[pasin]

@jianminzhao What is the latest strategy (e.g. using channels or docIDs filters) for setting up the replicator in tests with SG?

[jianminzhao]

This test is not included in PR validation. The validation failure must be due to something else, a.k.a flaky test.

[callumbirks]

This test is not included in PR validation. The validation failure must be due to something else, a.k.a flaky test.

I think it failed during the jenkins Windows or MacOS build because of an unused variable or something.

[callumbirks]

Test is now passing multiple times in a row without purging bucket or restarting SGW. Achieved this by purging doc at the end of test. Also now have temp user for test, as Pasin suggested. Test is now also future-proofed against the multiple-collection support in SGW as I have specified the document's collection in the PUT request.

[callumbirks]

Tests are passing, turns out it was due to some strange SGW bug. Possibly related to CBG-2487.

[callumbirks]

I've had to modify the assignUserChannel function to take a parameter of user because it was hardcoded "sguser"

[callumbirks]

I've temporarily set the test's tag to "SyncServerCollectionTEMP" until we decide what to do with the test while CBG-2487 is still causing flaky failures.