Austin Cory Bart

So far, natural conversation flow suggests the following terms: * groups * variants * participants * variations * alternates

This seems to be mostly working, except I don't think it works on files that are not separated into sections! That should be allowed, it should just treat it as...

I suppose the real model should actually be to have a way to do something like `get_student_tests(function_name: str)` that produces a list of the parameter/result pairs. Then you could have...

Additional static solution: inject a secret call just past the header of the function definition. ```python from cisc108 import assert_equal def func(a: int) -> int: __remember_calls('func', a) ; ... expected...

The proper thing to do, I suppose, is to trace this dynamically. In theory, this is what the `SandboxCallTracer` is meant to do, except it currently only captures function names....

Cool, this actually worked really easily on the Pedal side. Just need to implement it on the Skulpt side and it should be good to go!

Worked on the Skulpt side too. Now we just need to get a convenient assertion for it so instructors can demand at least N unique calls. Technically belongs in the...

Would this be `sys.args`? So you'd want to be able to provide a `sys.args` value conveniently? Because I think you can use the existing infrastructure to achieve this, but it's...

Conversation at SIGCSE is that likely we should put this on `run` and make it easy for the instructors to specify. Also probably provide a `set_sys_args` similar to how we...

Not sure where this left off, but I notice in the current version it seems like sys args leaks into the students' code execution. We should make sure that this...