Lol, you must live in a different world, lots of code is using regular bounds checked indexing and it is completely idiomatic. So idiomatic in fact that there is no nonpanicking slicing method for slices or strings, for example.
You don't need one, since you can use `.chars().nth()` (indexing a string needs you to iterate over it anyway, because utf8). Though I do feel that having a direct get() API would be nice
I have seen quite a bit of regular `[]` indexing, however it's almost always in very obvious situations, and most of the places where you might index in C/++ (iteration, or with a bounds check) I've seen people using iterators or the monadic way.
But yeah, it's not that unidiomatic. It's still used a lot, just not to the extent it's used elsewhere.
(For example, we have thrice as much indexing in our DOM binding generation python code (`components/script/dom/bindings/codegen`) in Servo than we do in all of the non-test rust files (`components/`) in the main repo combined)
