granit_parser/
char_traits.rs

1//! Holds functions to determine if a character belongs to a specific character set.
2
3/// Check whether the character is nil (`\0`).
4#[inline]
5#[must_use]
6pub fn is_z(c: char) -> bool {
7    c == '\0'
8}
9
10/// Check whether the character is a line break (`\r` or `\n`).
11#[inline]
12#[must_use]
13pub fn is_break(c: char) -> bool {
14    c == '\n' || c == '\r'
15}
16
17/// Check whether the character is nil or a line break (`\0`, `\r`, `\n`).
18#[inline]
19#[must_use]
20pub fn is_breakz(c: char) -> bool {
21    is_break(c) || is_z(c)
22}
23
24/// Check whether the character is a whitespace (` ` or `\t`).
25#[inline]
26#[must_use]
27pub fn is_blank(c: char) -> bool {
28    c == ' ' || c == '\t'
29}
30
31/// Check whether the character is nil, a line break, or whitespace.
32///
33/// `\0`, ` `, `\t`, `\n`, `\r`
34#[inline]
35#[must_use]
36pub fn is_blank_or_breakz(c: char) -> bool {
37    is_blank(c) || is_breakz(c)
38}
39
40/// Check whether the character is an ASCII digit.
41#[inline]
42#[must_use]
43pub fn is_digit(c: char) -> bool {
44    c.is_ascii_digit()
45}
46
47/// Check whether the character is an ASCII alphanumeric, `_` or `-`.
48///
49/// This is used for scanning tag handles and similar constructs.
50/// Note: This is slightly more permissive than YAML's `ns-word-char` (which excludes `_`).
51/// For strict `ns-word-char` compliance, use `is_word_char` instead.
52///
53/// Matches: `[0-9a-zA-Z_-]`
54#[inline]
55#[must_use]
56pub fn is_alpha(c: char) -> bool {
57    matches!(c, '0'..='9' | 'a'..='z' | 'A'..='Z' | '_' | '-')
58}
59
60/// Check whether the character is a hexadecimal character (case insensitive).
61#[inline]
62#[must_use]
63pub fn is_hex(c: char) -> bool {
64    c.is_ascii_digit() || ('a'..='f').contains(&c) || ('A'..='F').contains(&c)
65}
66
67/// Convert the hexadecimal digit to an integer.
68///
69/// # Panics
70/// Panics if `c` is not an ASCII hexadecimal digit.
71#[track_caller]
72#[inline]
73#[must_use]
74pub fn as_hex(c: char) -> u32 {
75    match c {
76        '0'..='9' => (c as u32) - ('0' as u32),
77        'a'..='f' => (c as u32) - ('a' as u32) + 10,
78        'A'..='F' => (c as u32) - ('A' as u32) + 10,
79        _ => unreachable!("as_hex called with a non-hexadecimal character"),
80    }
81}
82
83/// Check whether the character is a YAML flow character (one of `,[]{}`).
84#[inline]
85#[must_use]
86pub fn is_flow(c: char) -> bool {
87    matches!(c, ',' | '[' | ']' | '{' | '}')
88}
89
90/// Check whether the character is the BOM character.
91#[inline]
92#[must_use]
93pub fn is_bom(c: char) -> bool {
94    c == '\u{FEFF}'
95}
96
97/// Check whether the character is a YAML non-breaking character.
98#[inline]
99#[must_use]
100pub fn is_yaml_non_break(c: char) -> bool {
101    is_printable(c) && !is_break(c) && !is_bom(c)
102}
103
104/// Check whether the character is a YAML printable character (`c-printable`).
105#[inline]
106#[must_use]
107pub(crate) fn is_printable(c: char) -> bool {
108    matches!(
109        c as u32,
110        0x0009
111            | 0x000A
112            | 0x000D
113            | 0x0020..=0x007E
114            | 0x0085
115            | 0x00A0..=0xD7FF
116            | 0xE000..=0xFFFD
117            | 0x10000..=0x0010_FFFF
118    )
119}
120
121/// Check whether the character is NOT a YAML whitespace (` ` / `\t`).
122#[inline]
123#[must_use]
124pub fn is_yaml_non_space(c: char) -> bool {
125    is_yaml_non_break(c) && !is_blank(c)
126}
127
128/// Check whether the character is a valid YAML anchor name character.
129#[inline]
130#[must_use]
131pub fn is_anchor_char(c: char) -> bool {
132    is_yaml_non_space(c) && !is_flow(c) && !is_z(c)
133}
134
135/// Check whether the character is a valid YAML word character (`ns-word-char`).
136///
137/// Per YAML 1.2 spec: `ns-word-char ::= ns-dec-digit | ns-ascii-letter | "-"`
138///
139/// Matches: `[0-9a-zA-Z-]`
140#[inline]
141#[must_use]
142pub fn is_word_char(c: char) -> bool {
143    is_alpha(c) && c != '_'
144}
145
146/// Check whether the character is a valid URI character.
147#[inline]
148#[must_use]
149pub fn is_uri_char(c: char) -> bool {
150    is_word_char(c) || "#;/?:@&=+$,_.!~*\'()[]%".contains(c)
151}
152
153/// Check whether the character is a valid tag character.
154#[inline]
155#[must_use]
156pub fn is_tag_char(c: char) -> bool {
157    is_uri_char(c) && !is_flow(c) && c != '!'
158}
159
160#[cfg(test)]
161mod tests {
162    use super::*;
163
164    #[test]
165    fn printable_ranges_include_private_and_supplementary_planes() {
166        assert!(is_printable('\u{E000}'));
167        assert!(is_printable('\u{10FFFF}'));
168        assert!(is_yaml_non_break('\u{10000}'));
169        assert!(!is_yaml_non_break('\u{FEFF}'));
170        assert!(!is_yaml_non_break('\n'));
171    }
172
173    #[test]
174    fn word_uri_and_tag_character_sets_are_distinct() {
175        assert!(is_word_char('-'));
176        assert!(!is_word_char('_'));
177        assert!(is_uri_char('_'));
178        assert!(is_uri_char('%'));
179        assert!(!is_tag_char('!'));
180        assert!(!is_tag_char('['));
181    }
182}
granit_parser/char_traits.rs

granit_parser/
char_traits.rs
