reverse-string: add test that exposes unquoted variable problems

[glennj]

Addressing same problem as PR #377

---

Reviewer Resources:

Track Policies

[guygastineau]

Is this our own test? (ie. not from problem-specifications)

[glennj]

Is this our own test? (ie. not from problem-specifications)

Yes. It comes from seeing the same mistake over and over while mentoring.

I would not think this kind of problem exists in other languages, so I don't think it's worth submitting to the canonical data.

[IsaacG]

Thanks for adding this!

[guygastineau]

Right, but we don't get to increment the version past the canonical-data version I don't believe

[guygastineau]

The Dashboard doesn't seem to care, but what happens if the canonical-data.json is upgraded? We wont be able to tell that our exercise is behind the new information in the problem-specifications repository.

[guygastineau]

BTW, sorry I didn't see your response earlier. For the record I think this is a great test to add to this exercise for our track! 🤓

[glennj]

I've done it before, see https://tracks.exercism.io/bash/master (word-count, for example) -- the major and minor versions are the same, just the patch number is incremented.

[guygastineau]

Right, sometimes test semantics are changed and they just increment the patch number, so we will still be getting out of sync.

Unless I am missing something here.