Potential use of core::hint::unreachable_unchecked to avoid bounds checks for all users

[yescallop]

Consider the following code. It's used as a wrapper in one of my projects to avoid bounds checks in safe code.

#[inline]
fn chr(s: &[u8], b: u8) -> Option<usize> {
    memchr::memchr(b, s).map(|i| {
        if i >= s.len() {
            unsafe { core::hint::unreachable_unchecked() }
        }
        i
    })
}

I wonder if it's practical and sound to insert unreachable hints into memchr crate, so that all its users could get an increase in performance. It's just a tentative suggestion that needs more discussion :)

[BurntSushi]

Sorry, I don't quite understand the request here. I would say there are a few dimensions to it:

1. Do you have a real benchmark where eliding the check improves things? If so, I'd love to see it, because it is somewhat surprising to me that it would help much. In particular, memchr is always going to result in at least a pointer load and a function call, so a single branch in addition to that probably isn't adding much overhead.
2. More importantly, it's unclear to me why you have this conditional at all. Why not just elide if i >= s.len() { completely? Maybe what you're saying---without actually saying it---is that the conditional provides a way for other code to avoid bounds checks. If so, then I'd really appreciate more detail, particularly with respect to (1).
3. How would you go about moving the hint to the inside this crate?

[BurntSushi]

Closing due to inactivity and the next steps here are unclear to me.