Clean: Fix MSVC compiler warning.

[a858438680]

why do not change it to static assert?

[bjones1]

A good point. I like that, but then it means we can't (easily) test that this works. I found a blog post on this, but it seems like a lot of work for one test.

So, @jbaldwin, what do you prefer? The static assert would be ring_buffer() { static_assert(num_elements != 0, "num_elements cannot be zero"); } plus dropping the lines 8-12 in test_ring_buffer.cpp.

[jbaldwin]

Static assert is probably the correct way to solve this, pushing towards compile time errors is usually better. It's simple enough too we probably don't need to test that it fails to compile in a test IMO, but won't turn down a PR that does test it.

[bjones1]

Thanks @a858438680 for bringing this up! Here's the improved version per @jbaldwin's request.

[bjones1]

Note that the fedora-36 failure is a download problem, not a test failure.

[jbaldwin]

Retriggered the fed 36 build