Define scheme for running and publishing regular FediTest reports

[jernst]

Here is an idea that may be horrifically bad. (git submodule to the power of N):

• We create a git repo that only contains test results. Every time we "officially" (and we need to define what that is) run tests, we check the results into that repo as new versions of the results from last time.
• the feditest.org Hugo repo uses this test result repo as a git submodule so they are available as part of the feditest.org website. But not once, but it "mounts" the same repo in different places in the directory hierarchy, pointing to different commits. So we can have .../results/2024-06/ point to one version in the results repo, and ..../results/2024-07/ to another. That would sort of correspond to the structure of the data, but just about everybody warns of git submodule, and doing so many of them ...

And we can add one more: we could make the feditest framework repo, and the feditest-tests-fediverse (that contains the actual tests) as submodules of the test results repo. Then one could also know which versions of the tests were run with which version of the framework.

How crazy can one get with git submodules? Somebody stop me before I might actually do this ...

[jernst]

We won't publish regular reports, reasoning is in https://feditest.org/blog/2024-06-13-some-tentative-decisions/