Test code which uses reflect doesn't handle incomplete code well

[autarch]

If a student writes incomplete code that compiles and runs the tests, it's easy for the tests to panic. They probably need a lot more checks for nil all over the place.

[autarch]

@ribasushi - do you remember what code did this, by any chance? I'm trying to come up with an example to help replicate this. I vaguely remember it was something to do with implementing a func that had the right signature but maybe just returned nil?

[autarch]

Actually, I'm starting to wonder if this was just another instance of the bug fixed in stretchr/testify#622

[ribasushi]

@autarch almost certainly yes - the text of the panic() was nearly identical

[autarch]

Thanks for confirming. Once that PR is merged I think the bulk of these issues will go away. It just turns out that it's fairly easy for the exercise tests to tickle that particular bug.

[autarch]

Fixed by updating the deps for the class exercises.