I believe the fail() pattern came from jasmine down to jest in the first
instance, though I am not sure. In any event it leads to some strange
testing behavior in which you begin to re-invent the framework from the
inside out.

For this reason it seems the vite team has decided not to make a pass at
including it, since it is technically breaking containment even in the
context of jest testing.

I have done my best to take the time it needs to ascertain which of
these fails() should be asserts, and which should be using the standard
expectations api.

I have maintained the file-sequential execution style (which I will next
attempt to take away.) tests currently run through in ~99 seconds or so.
This is much longer than it needs to be, and so this is where the vitest
optimizations can really have an opportunity to shine.

The test speed is up roughly 10x but my expectation is that it should be
able to come lower than that. The beauty of the way test containment is
implemented here means that there is not much need for isolation - which
is good for simulating our use case because we make use of a lot of
singletons (though thankfully not TOO much shared state)

It was all relatively trivial with the exception of histories. I am
intensely anxious with the implementation that has been used for
document history, as it obscures control flow and does not look like it
will scale - especially if we were ever to want horizontal deployment.

All in all the tests are faster, and can tighten up development feedback
loops BUT the un-strict nature of our existing test runners means that
there are some inconsistencies in execution that I haven't yet uncovered

I still need a little guidance on what the area history behavior is supposed to be,
since it is not clear to me what approach is being used - it's only two
tests failing but it is indicative of my misunderstanding.

All tests resolve in about 8 seconds, which is quite nice

the vite config is also expressed in TS, so it should not be exempt
from linting

testing is quiiiiite noisy in testing

Thanks for adding the pagination tests!