Type of `sinceSeq` can be also a String besides an Integer

[mojito317]

Checklist

• [x] Tick to sign-off your agreement to the Developer Certificate of Origin (DCO) 1.1
• [x] Added tests for code changes or test/build only changes
• [x] Updated the change log file (CHANGES.md|CHANGELOG.md) or test/build only changes
• [x] Completed the PR template below:

Description

Fixes #526

Approach

seqs could be Integers in CouchDB 1.x and Strings in 2.x, so I changed all the places where they remained Integers to JsonElement. With this, both Integer and String can be used.

Security and Privacy

• No change

Testing

• This behavior can be tested with ITs well, so I just extended ITs with a plus .sinceSeq(...) call which is going to use a seq from a {source_db}/_changes request.

Monitoring and Logging

• No change

[mojito317]

Nearly there, as discussed offline I think the seq things should preferably be in their own tests.

Done in 79abd33. Lmk if you would like to see more.

don't change the existing API in the ReplicatorDocument model class

Done in 103eb3b and 0c4e3c8.

• changes entry

Done in 64770f4.

[mojito317]

+1 just a couple nits

Thanks for the thorough review! I fixed the nits in f3018dd.

[mojito317]

We should consider adding a negative test to assert that an error message is thrown when passing in an unexpected type.

An unexpected sinceSeq (e.g. a "true" or a "asd") will cause a crashing state in the replication and the replication will fail and timeout eventually in the waitForReplicatorToReachStatus method: https://github.com/cloudant/java-cloudant/blob/2c53b88a84906a3a47f844ce7c25078e11baee46/cloudant-client/src/test/java/com/cloudant/tests/util/Utils.java#L142

Meanwhile, the replication document will disappear from the _replicator database.

Other types cannot be added, because compilation would fail.

[mojito317]

Thanks for the reviews @ricellis and @emlaver! I'm about to merge this!