src/byteset/scalar.rs

// This is adapted from `fallback.rs` from rust-memchr. It's modified to return
// the 'inverse' query of memchr, e.g. finding the first byte not in the provided
// set. This is simple for the 1-byte case.

use core::cmp;
use core::usize;

#[cfg(target_pointer_width = "32")]
const USIZE_BYTES: usize = 4;

#[cfg(target_pointer_width = "64")]
const USIZE_BYTES: usize = 8;

// The number of bytes to loop at in one iteration of memchr/memrchr.
const LOOP_SIZE: usize = 2 * USIZE_BYTES;

/// Repeat the given byte into a word size number. That is, every 8 bits
/// is equivalent to the given byte. For example, if `b` is `\x4E` or
/// `01001110` in binary, then the returned value on a 32-bit system would be:
/// `01001110_01001110_01001110_01001110`.
#[inline(always)]
fn repeat_byte(b: u8) -> usize {
    (b as usize) * (usize::MAX / 255)
}

pub fn inv_memchr(n1: u8, haystack: &[u8]) -> Option<usize> {
    let vn1 = repeat_byte(n1);
    let confirm = |byte| byte != n1;
    let loop_size = cmp::min(LOOP_SIZE, haystack.len());
    let align = USIZE_BYTES - 1;
    let start_ptr = haystack.as_ptr();
    let end_ptr = haystack[haystack.len()..].as_ptr();
    let mut ptr = start_ptr;

    unsafe {
        if haystack.len() < USIZE_BYTES {
            return forward_search(start_ptr, end_ptr, ptr, confirm);
        }

        let chunk = read_unaligned_usize(ptr);
        if (chunk ^ vn1) != 0 {
            return forward_search(start_ptr, end_ptr, ptr, confirm);
        }

        ptr = ptr.add(USIZE_BYTES - (start_ptr as usize & align));
        debug_assert!(ptr > start_ptr);
        debug_assert!(end_ptr.sub(USIZE_BYTES) >= start_ptr);
        while loop_size == LOOP_SIZE && ptr <= end_ptr.sub(loop_size) {
            debug_assert_eq!(0, (ptr as usize) % USIZE_BYTES);

            let a = *(ptr as *const usize);
            let b = *(ptr.add(USIZE_BYTES) as *const usize);
            let eqa = (a ^ vn1) != 0;
            let eqb = (b ^ vn1) != 0;
            if eqa || eqb {
                break;
            }
            ptr = ptr.add(LOOP_SIZE);
        }
        forward_search(start_ptr, end_ptr, ptr, confirm)
    }
}

/// Return the last index not matching the byte `x` in `text`.
pub fn inv_memrchr(n1: u8, haystack: &[u8]) -> Option<usize> {
    let vn1 = repeat_byte(n1);
    let confirm = |byte| byte != n1;
    let loop_size = cmp::min(LOOP_SIZE, haystack.len());
    let align = USIZE_BYTES - 1;
    let start_ptr = haystack.as_ptr();
    let end_ptr = haystack[haystack.len()..].as_ptr();
    let mut ptr = end_ptr;

    unsafe {
        if haystack.len() < USIZE_BYTES {
            return reverse_search(start_ptr, end_ptr, ptr, confirm);
        }

        let chunk = read_unaligned_usize(ptr.sub(USIZE_BYTES));
        if (chunk ^ vn1) != 0 {
            return reverse_search(start_ptr, end_ptr, ptr, confirm);
        }

        ptr = (end_ptr as usize & !align) as *const u8;
        debug_assert!(start_ptr <= ptr && ptr <= end_ptr);
        while loop_size == LOOP_SIZE && ptr >= start_ptr.add(loop_size) {
            debug_assert_eq!(0, (ptr as usize) % USIZE_BYTES);

            let a = *(ptr.sub(2 * USIZE_BYTES) as *const usize);
            let b = *(ptr.sub(1 * USIZE_BYTES) as *const usize);
            let eqa = (a ^ vn1) != 0;
            let eqb = (b ^ vn1) != 0;
            if eqa || eqb {
                break;
            }
            ptr = ptr.sub(loop_size);
        }
        reverse_search(start_ptr, end_ptr, ptr, confirm)
    }
}

#[inline(always)]
unsafe fn forward_search<F: Fn(u8) -> bool>(
    start_ptr: *const u8,
    end_ptr: *const u8,
    mut ptr: *const u8,
    confirm: F,
) -> Option<usize> {
    debug_assert!(start_ptr <= ptr);
    debug_assert!(ptr <= end_ptr);

    while ptr < end_ptr {
        if confirm(*ptr) {
            return Some(sub(ptr, start_ptr));
        }
        ptr = ptr.offset(1);
    }
    None
}

#[inline(always)]
unsafe fn reverse_search<F: Fn(u8) -> bool>(
    start_ptr: *const u8,
    end_ptr: *const u8,
    mut ptr: *const u8,
    confirm: F,
) -> Option<usize> {
    debug_assert!(start_ptr <= ptr);
    debug_assert!(ptr <= end_ptr);

    while ptr > start_ptr {
        ptr = ptr.offset(-1);
        if confirm(*ptr) {
            return Some(sub(ptr, start_ptr));
        }
    }
    None
}

unsafe fn read_unaligned_usize(ptr: *const u8) -> usize {
    (ptr as *const usize).read_unaligned()
}

/// Subtract `b` from `a` and return the difference. `a` should be greater than
/// or equal to `b`.
fn sub(a: *const u8, b: *const u8) -> usize {
    debug_assert!(a >= b);
    (a as usize) - (b as usize)
}

/// Safe wrapper around `forward_search`
#[inline]
pub(crate) fn forward_search_bytes<F: Fn(u8) -> bool>(
    s: &[u8],
    confirm: F,
) -> Option<usize> {
    unsafe {
        let start = s.as_ptr();
        let end = start.add(s.len());
        forward_search(start, end, start, confirm)
    }
}

/// Safe wrapper around `reverse_search`
#[inline]
pub(crate) fn reverse_search_bytes<F: Fn(u8) -> bool>(
    s: &[u8],
    confirm: F,
) -> Option<usize> {
    unsafe {
        let start = s.as_ptr();
        let end = start.add(s.len());
        reverse_search(start, end, end, confirm)
    }
}

#[cfg(test)]
mod tests {
    use super::{inv_memchr, inv_memrchr};
    // search string, search byte, inv_memchr result, inv_memrchr result.
    // these are expanded into a much larger set of tests in build_tests
    const TESTS: &[(&[u8], u8, usize, usize)] = &[
        (b"z", b'a', 0, 0),
        (b"zz", b'a', 0, 1),
        (b"aza", b'a', 1, 1),
        (b"zaz", b'a', 0, 2),
        (b"zza", b'a', 0, 1),
        (b"zaa", b'a', 0, 0),
        (b"zzz", b'a', 0, 2),
    ];

    type TestCase = (Vec<u8>, u8, Option<(usize, usize)>);

    fn build_tests() -> Vec<TestCase> {
        let mut result = vec![];
        for &(search, byte, fwd_pos, rev_pos) in TESTS {
            result.push((search.to_vec(), byte, Some((fwd_pos, rev_pos))));
            for i in 1..515 {
                // add a bunch of copies of the search byte to the end.
                let mut suffixed: Vec<u8> = search.into();
                suffixed.extend(std::iter::repeat(byte).take(i));
                result.push((suffixed, byte, Some((fwd_pos, rev_pos))));

                // add a bunch of copies of the search byte to the start.
                let mut prefixed: Vec<u8> =
                    std::iter::repeat(byte).take(i).collect();
                prefixed.extend(search);
                result.push((
                    prefixed,
                    byte,
                    Some((fwd_pos + i, rev_pos + i)),
                ));

                // add a bunch of copies of the search byte to both ends.
                let mut surrounded: Vec<u8> =
                    std::iter::repeat(byte).take(i).collect();
                surrounded.extend(search);
                surrounded.extend(std::iter::repeat(byte).take(i));
                result.push((
                    surrounded,
                    byte,
                    Some((fwd_pos + i, rev_pos + i)),
                ));
            }
        }

        // build non-matching tests for several sizes
        for i in 0..515 {
            result.push((
                std::iter::repeat(b'\0').take(i).collect(),
                b'\0',
                None,
            ));
        }

        result
    }

    #[test]
    fn test_inv_memchr() {
        use crate::{ByteSlice, B};
        for (search, byte, matching) in build_tests() {
            assert_eq!(
                inv_memchr(byte, &search),
                matching.map(|m| m.0),
                "inv_memchr when searching for {:?} in {:?}",
                byte as char,
                // better printing
                B(&search).as_bstr(),
            );
            assert_eq!(
                inv_memrchr(byte, &search),
                matching.map(|m| m.1),
                "inv_memrchr when searching for {:?} in {:?}",
                byte as char,
                // better printing
                B(&search).as_bstr(),
            );
            // Test a rather large number off offsets for potential alignment issues
            for offset in 1..130 {
                if offset >= search.len() {
                    break;
                }
                // If this would cause us to shift the results off the end, skip
                // it so that we don't have to recompute them.
                if let Some((f, r)) = matching {
                    if offset > f || offset > r {
                        break;
                    }
                }
                let realigned = &search[offset..];

                let forward_pos = matching.map(|m| m.0 - offset);
                let reverse_pos = matching.map(|m| m.1 - offset);

                assert_eq!(
                    inv_memchr(byte, &realigned),
                    forward_pos,
                    "inv_memchr when searching (realigned by {}) for {:?} in {:?}",
                    offset,
                    byte as char,
                    realigned.as_bstr(),
                );
                assert_eq!(
                    inv_memrchr(byte, &realigned),
                    reverse_pos,
                    "inv_memrchr when searching (realigned by {}) for {:?} in {:?}",
                    offset,
                    byte as char,
                    realigned.as_bstr(),
                );
            }
        }
    }
}