TestPrivateGroupsManualDecrypt is flaky about go-ssb

[KyleMaas]

I think a lot of this may have been fixed by the Publish stuff becoming synchronous. It means it has to slow down the test a little bit and do things in order, which makes it more likely to have been fully processed by the time it tries to replicate the data into the second bot. I'm sure the problem's still there, but it's likely even more rare (and thus flaky) now than before.

[KyleMaas]

So...this one's going to be a little tricky to debug. I've run it a little over 100 times now with zero failures. There are plenty of possibilities for race conditions, several of which we've already solved. So this might or might not already be fixed. One way or another, it's rare enough that it's going to be hard to know for sure if and when it has actually been fixed.