syn/
whitespace.rs

1// SPDX-License-Identifier: Apache-2.0 OR MIT
2
3pub(crate) fn skip(mut s: &str) -> &str {
4    'skip: while !s.is_empty() {
5        let byte = s.as_bytes()[0];
6        if byte == b'/' {
7            if s.starts_with("//")
8                && (!s.starts_with("///") || s.starts_with("////"))
9                && !s.starts_with("//!")
10            {
11                if let Some(i) = s.find('\n') {
12                    s = &s[i + 1..];
13                    continue;
14                } else {
15                    return "";
16                }
17            } else if s.starts_with("/**/") {
18                s = &s[4..];
19                continue;
20            } else if s.starts_with("/*")
21                && (!s.starts_with("/**") || s.starts_with("/***"))
22                && !s.starts_with("/*!")
23            {
24                let mut depth = 0;
25                let bytes = s.as_bytes();
26                let mut i = 0;
27                let upper = bytes.len() - 1;
28                while i < upper {
29                    if bytes[i] == b'/' && bytes[i + 1] == b'*' {
30                        depth += 1;
31                        i += 1; // eat '*'
32                    } else if bytes[i] == b'*' && bytes[i + 1] == b'/' {
33                        depth -= 1;
34                        if depth == 0 {
35                            s = &s[i + 2..];
36                            continue 'skip;
37                        }
38                        i += 1; // eat '/'
39                    }
40                    i += 1;
41                }
42                return s;
43            }
44        }
45        match byte {
46            b' ' | 0x09..=0x0D => {
47                s = &s[1..];
48                continue;
49            }
50            b if b <= 0x7F => {}
51            _ => {
52                let ch = s.chars().next().unwrap();
53                if is_whitespace(ch) {
54                    s = &s[ch.len_utf8()..];
55                    continue;
56                }
57            }
58        }
59        return s;
60    }
61    s
62}
63
64fn is_whitespace(ch: char) -> bool {
65    // Rust treats left-to-right mark and right-to-left mark as whitespace
66    ch.is_whitespace() || ch == '\u{200e}' || ch == '\u{200f}'
67}
syn/whitespace.rs

syn/
whitespace.rs
