A sequence of RaCKet expressions that are meant to be Loaded (as opposed to a module to be required).

Is it done this way for historical reasons, or are there advantages to loading vs. requiring?

For the core test suite, it’s historical. There may also be advantages to the load style in that context, since the test suite is mostly scripting runs of the interpreter and testing things that are below the layer of modules. But whether there are any advantages, the load style has been a good enough fit to avoid a rewrite/reorganization of the test suite.

All right - my first pull request was accepted! (I don’t count the typo in the DrRacket easter egg :) ). Looking forward to many more contributions in the future. I suppose it helped that much of the code was Matthew’s since it was a bit better than my original implementation, but I’m still counting it because the change was my idea, and I did make a minor improvement :)