chore: fix remote server test for duplicate server add issue

[gmaclennan]

For #1200 I skipped running the new failing test on a remote server because the test setup was failing. The underlying issue is that there is a different race condition bug in the server when two clients try to add the same project to the server at the same time, which is low-priority to fix and I'm tracking in this issue: digidem/comapeo-cloud#66

This PR updates the test to add the duplicate server to the project sequentially. The underlying issue was not adding a duplicate server at exactly the same time, but two devices adding the same server while not connected / syncing with each other, so this updated test should catch the issue.

Although this PR only updates the test, I have reverted the fix, then added the test as a separate commit, to verify that the test actually fails prior to the fix still.

I recommend reviewing with the "Hide whitespace" option in Github, to help focus on the lines that have changed.

[RangerMauve]

LGTM.

Do you know why this change works where the parallel case failed? If that's an issue we'll likely encounter it eventually. (race condition in calling get project while another get project is in progress?)

[gmaclennan]

The issue causing the original parallel test failure is digidem/comapeo-cloud#66