Update ensure_filter_type_is_valid upper bound enum.

[bekadavis9]

Update the upper bound in ensure_filter_type_is_valid to reflect the highest-possible filter_type enum.

---

[sc-65265]

---

TYPE: NO_HISTORY
DESC: Update ensure_filter_type_is_valid upper bound enum.

[rroelke]

I definitely don't love seeing an un-named constant here, it will be tricky to remember this if another filter type is added.

Maybe we can add a variant to the enum class FilterType which is INTERNAL_FILTER_COUNT or something? This definitely works with enums which don't have values assigned, but maybe it will work here anyway.

[rroelke]

Another way to ensure that this does not break in the future would be to claim that the type is invalid if and only if filter_type_str returns empty string. That buys some robustness against someone forgetting to update both of those functions.

[bekadavis9]

100% agree. It's been discussed in the past that all of these ensure_<enum_type>_is_valid methods should check on some constant so programmers aren't responsible for hopefully remembering to update them (if they even know about them). This is certainly worth a larger discussion offline. Note that this PR is just a one-off update after I noticed this method was out-of-date during a separate discussion.

[rroelke]

Let's add a quick test here which just checks if (is_filter_type_valid(i)) { CHECK(filter_type_str(i) != constants::empty_string); }) and also the other direction. I think that's a good enough way of adding future robustness.