use crate::unicode_key_table::UNICODE_KEY_RANGES;
#[inline(always)]
fn category(code_point: u32) -> u32 {
let mut low = 0usize;
let mut high = UNICODE_KEY_RANGES.len() / 3;
while low < high {
let mid = (low + high) / 2;
let offset = mid * 3;
if code_point < UNICODE_KEY_RANGES[offset] {
high = mid;
} else if code_point > UNICODE_KEY_RANGES[offset + 1] {
low = mid + 1;
} else {
return UNICODE_KEY_RANGES[offset + 2];
}
}
0
}
#[inline(always)]
pub(crate) fn is_initial(c: char) -> bool {
if c.is_ascii() {
c.is_ascii_alphanumeric() || c == '_'
} else {
category(c as u32) & 5 != 0
}
}
#[inline(always)]
pub(crate) fn is_character(c: char) -> bool {
if c.is_ascii() {
is_initial(c) || matches!(c, ':' | '-' | '*')
} else {
c == '·' || category(c as u32) != 0
}
}
#[inline(always)]
pub(crate) fn is_valid_unquoted(raw: &str, allow_slash: bool) -> bool {
let bytes = raw.as_bytes();
let Some(&first) = bytes.first() else {
return false;
};
let mut i = if first.is_ascii() {
if !(first.is_ascii_alphanumeric() || first == b'_') {
return false;
}
let mut non_ascii = None;
for (offset, &b) in bytes[1..].iter().enumerate() {
if !b.is_ascii() {
non_ascii = Some(offset + 1);
break;
}
if !(b.is_ascii_alphanumeric()
|| matches!(b, b'_' | b':' | b'-' | b'*')
|| allow_slash && b == b'/')
{
return false;
}
}
let Some(offset) = non_ascii else {
return true;
};
offset
} else {
let c = raw.chars().next().unwrap();
if !is_initial(c) {
return false;
}
c.len_utf8()
};
while i < bytes.len() {
if bytes[i].is_ascii() {
let b = bytes[i];
if !(b.is_ascii_alphanumeric()
|| matches!(b, b'_' | b':' | b'-' | b'*')
|| allow_slash && b == b'/')
{
return false;
}
i += 1;
} else {
let c = raw[i..].chars().next().unwrap();
if !(is_character(c) || allow_slash && c == '/') {
return false;
}
i += c.len_utf8();
}
}
true
}