#![no_std]
#![feature(stdsimd)]
#![feature(doc_cfg)]
pub mod lemire;
#[cfg(test)]
mod tests {
macro_rules! create_tests {
($test_function:ident) => {
assert_eq!($test_function(&[0xc0, 0x80]), false);
assert_eq!($test_function(&[0xc0, 0xae]), false);
assert_eq!($test_function(&[0xe0, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xe0, 0x80, 0xaf]), false);
assert_eq!($test_function(&[0xe0, 0x81, 0x81]), false);
assert_eq!($test_function(&[0xf0, 0x82, 0x82, 0xac]), false);
assert_eq!($test_function(&[0xf4, 0x90, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xED, 0xA0, 0x80]), false);
assert_eq!($test_function(&[0xED, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xC2, 0x80]), true);
assert_eq!($test_function(&[0xDF, 0xBF]), true);
assert_eq!($test_function(&[0xE0, 0xA0, 0x80]), true);
assert_eq!($test_function(&[0xED, 0x9F, 0xBF]), true);
assert_eq!($test_function(&[0xEE, 0x80, 0x80]), true);
assert_eq!($test_function(&[0xEF, 0xBF, 0xBF]), true);
assert_eq!($test_function(&[0xF0, 0x90, 0x80, 0x80]), true);
assert_eq!($test_function(&[0xF4, 0x8F, 0xBF, 0xBF]), true);
assert_eq!($test_function("κόσμε".as_bytes()), true);
assert_eq!($test_function(&[0]), true);
assert_eq!($test_function(&[0xC2, 0x80]), true);
assert_eq!($test_function(&[0xE0, 0xA0, 0x80]), true);
assert_eq!($test_function(&[0xF0, 0x90, 0x80, 0x80]), true);
assert_eq!($test_function(&[0xF8, 0x88, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xFC, 0x84, 0x80, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0x7F]), true);
assert_eq!($test_function(&[0xDF, 0xBF]), true);
assert_eq!($test_function(&[0xEF, 0xBF, 0xBF]), true);
assert_eq!($test_function(&[0xF7, 0xBF, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xFB, 0xBF, 0xBF, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xFD, 0xBF, 0xBF, 0xBF, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xED, 0x9F, 0xBF]), true);
assert_eq!($test_function(&[0xEE, 0x80, 0x80]), true);
assert_eq!($test_function(&[0xEF, 0xBF, 0xBD]), true);
assert_eq!($test_function(&[0xF4, 0x8F, 0xBF, 0xBF]), true);
assert_eq!($test_function(&[0xF4, 0x90, 0x80, 0x80]), false);
assert_eq!($test_function(&[0x80]), false);
assert_eq!($test_function(&[0xbf]), false);
assert_eq!($test_function(&[0x80, 0xBF]), false);
assert_eq!($test_function(&[0x80, 0xBF, 0x80]), false);
assert_eq!($test_function(&[0x80, 0xBF, 0x80, 0xBF]), false);
assert_eq!($test_function(&[0x80, 0xBF, 0x80, 0xBF, 0x80]), false);
assert_eq!($test_function(&[0x80, 0xBF, 0x80, 0xBF, 0x80, 0xBF]), false);
assert_eq!($test_function(&[0x80, 0xBF, 0x80, 0xBF, 0x80, 0xBF, 0x80]), false);
#[cfg_attr(rustfmt, rustfmt_skip)]
let continuation_bytes = [
0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87,
0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F,
0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97,
0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9E, 0x9F,
0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7,
0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF,
0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7,
0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF,
];
assert_eq!($test_function(&continuation_bytes), false);
for &b in continuation_bytes.iter() {
assert_eq!($test_function(&[b]), false);
}
#[cfg_attr(rustfmt, rustfmt_skip)]
let lonely_start_characters_2 = [
0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7,
0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF,
0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7,
0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF,
];
assert_eq!($test_function(&lonely_start_characters_2), false);
for &b in &lonely_start_characters_2 {
assert_eq!($test_function(&[b]), false);
}
#[cfg_attr(rustfmt, rustfmt_skip)]
let lonely_start_characters_3 = [
0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7,
0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF,
];
assert_eq!($test_function(&lonely_start_characters_3), false);
for &b in &lonely_start_characters_3 {
assert_eq!($test_function(&[b]), false);
}
#[cfg_attr(rustfmt, rustfmt_skip)]
let lonely_start_characters_4 = [
0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7,
];
assert_eq!($test_function(&lonely_start_characters_4), false);
for &b in &lonely_start_characters_4 {
assert_eq!($test_function(&[b]), false);
}
let lonely_start_characters_5 = [0xF8, 0xF9, 0xFA, 0xFB];
assert_eq!($test_function(&lonely_start_characters_5), false);
for &b in &lonely_start_characters_5 {
assert_eq!($test_function(&[b]), false);
}
let lonely_start_characters_6 = [0xFC, 0xFD];
assert_eq!($test_function(&lonely_start_characters_6), false);
for &b in &lonely_start_characters_6 {
assert_eq!($test_function(&[b]), false);
}
assert_eq!($test_function(&[0xC0]), false);
assert_eq!($test_function(&[0xE0, 0x80]), false);
assert_eq!($test_function(&[0xF0, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xF8, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xFC, 0x80, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xDF]), false);
assert_eq!($test_function(&[0xEF, 0xBF]), false);
assert_eq!($test_function(&[0xF7, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xFB, 0xBF, 0xBF, 0xBF]), false);
assert_eq!($test_function(&[0xFD, 0xBF, 0xBF, 0xBF, 0xBF]), false);
#[cfg_attr(rustfmt, rustfmt_skip)]
let incomplete = [
0xC0,
0xE0, 0x80,
0xF0, 0x80, 0x80,
0xF8, 0x80, 0x80, 0x80,
0xFC, 0x80, 0x80, 0x80, 0x80,
0xDF,
0xEF, 0xBF,
0xF7, 0xBF, 0xBF,
0xFB, 0xBF, 0xBF, 0xBF,
0xFD, 0xBF, 0xBF, 0xBF, 0xBF];
assert_eq!($test_function(&incomplete), false);
assert_eq!($test_function(&[0xFE]), false);
assert_eq!($test_function(&[0xFF]), false);
assert_eq!($test_function(&[0xFE, 0xFE, 0xFF, 0xFF]), false);
assert_eq!($test_function(&[0xC0, 0xAF]), false);
assert_eq!($test_function(&[0xE0, 0x80, 0xAF]), false);
assert_eq!($test_function(&[0xF0, 0x80, 0x80, 0xAF]), false);
assert_eq!($test_function(&[0xF8, 0x80, 0x80, 0x80, 0xAF]), false);
assert_eq!($test_function(&[0xFC, 0x80, 0x80, 0x80, 0x80, 0xAF]), false);
assert_eq!($test_function(&[0xC0, 0x80]), false);
assert_eq!($test_function(&[0xE0, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xF0, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xF8, 0x80, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xFC, 0x80, 0x80, 0x80, 0x80, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xa0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xad, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xae, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xaf, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xb0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xbe, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xbf, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xa0, 0x80, 0xed, 0xb0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xa0, 0x80, 0xed, 0xbf, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xad, 0xbf, 0xed, 0xb0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xad, 0xbf, 0xed, 0xbf, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xae, 0x80, 0xed, 0xb0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xae, 0x80, 0xed, 0xbf, 0xbf]), false);
assert_eq!($test_function(&[0xed, 0xaf, 0xbf, 0xed, 0xb0, 0x80]), false);
assert_eq!($test_function(&[0xed, 0xaf, 0xbf, 0xed, 0xbf, 0xbf]), false);
}
}
#[test]
fn test_lemire_avx() {
use super::lemire::avx::is_utf8;
create_tests!(is_utf8);
}
#[test]
fn test_lemire_avx_ascii() {
use super::lemire::avx::is_utf8_ascii_path;
create_tests!(is_utf8_ascii_path);
}
#[test]
fn test_lemire_sse() {
use super::lemire::sse::is_utf8;
create_tests!(is_utf8);
}
}