1 // Copyright 2015 The Servo Project Developers. See the
2 // COPYRIGHT file at the top-level directory of this distribution.
3 //
4 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
5 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
6 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
7 // option. This file may not be copied, modified, or distributed
8 // except according to those terms.
9
10 //! Accessor for `Bidi_Class` property from Unicode Character Database (UCD)
11
12 mod tables;
13
14 pub use self::tables::{BidiClass, UNICODE_VERSION};
15 #[cfg(feature = "hardcoded-data")]
16 use core::char;
17 #[cfg(feature = "hardcoded-data")]
18 use core::cmp::Ordering::{Equal, Greater, Less};
19
20 #[cfg(feature = "hardcoded-data")]
21 use self::tables::bidi_class_table;
22 use crate::data_source::BidiMatchedOpeningBracket;
23 use crate::BidiClass::*;
24 #[cfg(feature = "hardcoded-data")]
25 use crate::BidiDataSource;
26 /// Hardcoded Bidi data that ships with the unicode-bidi crate.
27 ///
28 /// This can be enabled with the default `hardcoded-data` Cargo feature.
29 #[cfg(feature = "hardcoded-data")]
30 pub struct HardcodedBidiData;
31
32 #[cfg(feature = "hardcoded-data")]
33 impl BidiDataSource for HardcodedBidiData {
bidi_class(&self, c: char) -> BidiClass34 fn bidi_class(&self, c: char) -> BidiClass {
35 bsearch_range_value_table(c, bidi_class_table)
36 }
37 }
38
39 /// Find the `BidiClass` of a single char.
40 #[cfg(feature = "hardcoded-data")]
bidi_class(c: char) -> BidiClass41 pub fn bidi_class(c: char) -> BidiClass {
42 bsearch_range_value_table(c, bidi_class_table)
43 }
44
45 /// If this character is a bracket according to BidiBrackets.txt,
46 /// return the corresponding *normalized* *opening bracket* of the pair,
47 /// and whether or not it itself is an opening bracket.
bidi_matched_opening_bracket(c: char) -> Option<BidiMatchedOpeningBracket>48 pub(crate) fn bidi_matched_opening_bracket(c: char) -> Option<BidiMatchedOpeningBracket> {
49 for pair in self::tables::bidi_pairs_table {
50 if pair.0 == c || pair.1 == c {
51 let skeleton = pair.2.unwrap_or(pair.0);
52 return Some(BidiMatchedOpeningBracket {
53 opening: skeleton,
54 is_open: pair.0 == c,
55 });
56 }
57 }
58 None
59 }
60
is_rtl(bidi_class: BidiClass) -> bool61 pub fn is_rtl(bidi_class: BidiClass) -> bool {
62 match bidi_class {
63 RLE | RLO | RLI => true,
64 _ => false,
65 }
66 }
67
68 #[cfg(feature = "hardcoded-data")]
bsearch_range_value_table(c: char, r: &'static [(char, char, BidiClass)]) -> BidiClass69 fn bsearch_range_value_table(c: char, r: &'static [(char, char, BidiClass)]) -> BidiClass {
70 match r.binary_search_by(|&(lo, hi, _)| {
71 if lo <= c && c <= hi {
72 Equal
73 } else if hi < c {
74 Less
75 } else {
76 Greater
77 }
78 }) {
79 Ok(idx) => {
80 let (_, _, cat) = r[idx];
81 cat
82 }
83 // UCD/extracted/DerivedBidiClass.txt: "All code points not explicitly listed
84 // for Bidi_Class have the value Left_To_Right (L)."
85 Err(_) => L,
86 }
87 }
88
89 #[cfg(all(test, feature = "hardcoded-data"))]
90 mod tests {
91 use super::*;
92
93 #[test]
test_ascii()94 fn test_ascii() {
95 assert_eq!(bidi_class('\u{0000}'), BN);
96 assert_eq!(bidi_class('\u{0040}'), ON);
97 assert_eq!(bidi_class('\u{0041}'), L);
98 assert_eq!(bidi_class('\u{0062}'), L);
99 assert_eq!(bidi_class('\u{007F}'), BN);
100 }
101
102 #[test]
test_bmp()103 fn test_bmp() {
104 // Hebrew
105 assert_eq!(bidi_class('\u{0590}'), R);
106 assert_eq!(bidi_class('\u{05D0}'), R);
107 assert_eq!(bidi_class('\u{05D1}'), R);
108 assert_eq!(bidi_class('\u{05FF}'), R);
109
110 // Arabic
111 assert_eq!(bidi_class('\u{0600}'), AN);
112 assert_eq!(bidi_class('\u{0627}'), AL);
113 assert_eq!(bidi_class('\u{07BF}'), AL);
114
115 // Default R + Arabic Extras
116 assert_eq!(bidi_class('\u{07C0}'), R);
117 assert_eq!(bidi_class('\u{085F}'), R);
118 assert_eq!(bidi_class('\u{0860}'), AL);
119 assert_eq!(bidi_class('\u{0870}'), AL);
120 assert_eq!(bidi_class('\u{089F}'), NSM);
121 assert_eq!(bidi_class('\u{08A0}'), AL);
122 assert_eq!(bidi_class('\u{089F}'), NSM);
123 assert_eq!(bidi_class('\u{08FF}'), NSM);
124
125 // Default ET
126 assert_eq!(bidi_class('\u{20A0}'), ET);
127 assert_eq!(bidi_class('\u{20CF}'), ET);
128
129 // Arabic Presentation Forms
130 assert_eq!(bidi_class('\u{FB1D}'), R);
131 assert_eq!(bidi_class('\u{FB4F}'), R);
132 assert_eq!(bidi_class('\u{FB50}'), AL);
133 assert_eq!(bidi_class('\u{FDCF}'), ON);
134 assert_eq!(bidi_class('\u{FDF0}'), AL);
135 assert_eq!(bidi_class('\u{FDFF}'), ON);
136 assert_eq!(bidi_class('\u{FE70}'), AL);
137 assert_eq!(bidi_class('\u{FEFE}'), AL);
138 assert_eq!(bidi_class('\u{FEFF}'), BN);
139
140 // noncharacters
141 assert_eq!(bidi_class('\u{FDD0}'), L);
142 assert_eq!(bidi_class('\u{FDD1}'), L);
143 assert_eq!(bidi_class('\u{FDEE}'), L);
144 assert_eq!(bidi_class('\u{FDEF}'), L);
145 assert_eq!(bidi_class('\u{FFFE}'), L);
146 assert_eq!(bidi_class('\u{FFFF}'), L);
147 }
148
149 #[test]
test_smp()150 fn test_smp() {
151 // Default AL + R
152 assert_eq!(bidi_class('\u{10800}'), R);
153 assert_eq!(bidi_class('\u{10FFF}'), R);
154 assert_eq!(bidi_class('\u{1E800}'), R);
155 assert_eq!(bidi_class('\u{1EDFF}'), R);
156 assert_eq!(bidi_class('\u{1EE00}'), AL);
157 assert_eq!(bidi_class('\u{1EEFF}'), AL);
158 assert_eq!(bidi_class('\u{1EF00}'), R);
159 assert_eq!(bidi_class('\u{1EFFF}'), R);
160 }
161
162 #[test]
test_unassigned_planes()163 fn test_unassigned_planes() {
164 assert_eq!(bidi_class('\u{30000}'), L);
165 assert_eq!(bidi_class('\u{40000}'), L);
166 assert_eq!(bidi_class('\u{50000}'), L);
167 assert_eq!(bidi_class('\u{60000}'), L);
168 assert_eq!(bidi_class('\u{70000}'), L);
169 assert_eq!(bidi_class('\u{80000}'), L);
170 assert_eq!(bidi_class('\u{90000}'), L);
171 assert_eq!(bidi_class('\u{a0000}'), L);
172 }
173 }
174