]>
Commit | Line | Data |
---|---|---|
abe05a73 XL |
1 | // Copyright 2012-2015 The Rust Project Developers. See the COPYRIGHT |
2 | // file at the top-level directory of this distribution and at | |
3 | // http://rust-lang.org/COPYRIGHT. | |
4 | // | |
5 | // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or | |
6 | // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license | |
7 | // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your | |
8 | // option. This file may not be copied, modified, or distributed | |
9 | // except according to those terms. | |
10 | ||
11 | //! Functions for computing canonical and compatible decompositions for Unicode characters. | |
f035d41b | 12 | use crate::lookups::{ |
6a06907d XL |
13 | canonical_fully_decomposed, cjk_compat_variants_fully_decomposed, |
14 | compatibility_fully_decomposed, composition_table, | |
f035d41b XL |
15 | }; |
16 | ||
17 | use core::{char, ops::FnMut}; | |
abe05a73 XL |
18 | |
19 | /// Compute canonical Unicode decomposition for character. | |
20 | /// See [Unicode Standard Annex #15](http://www.unicode.org/reports/tr15/) | |
21 | /// for more information. | |
8faf50e0 | 22 | #[inline] |
f035d41b XL |
23 | pub fn decompose_canonical<F>(c: char, emit_char: F) |
24 | where | |
25 | F: FnMut(char), | |
26 | { | |
dfeec247 | 27 | decompose(c, canonical_fully_decomposed, emit_char) |
8faf50e0 | 28 | } |
abe05a73 XL |
29 | |
30 | /// Compute canonical or compatible Unicode decomposition for character. | |
31 | /// See [Unicode Standard Annex #15](http://www.unicode.org/reports/tr15/) | |
32 | /// for more information. | |
8faf50e0 XL |
33 | #[inline] |
34 | pub fn decompose_compatible<F: FnMut(char)>(c: char, emit_char: F) { | |
f035d41b XL |
35 | let decompose_char = |
36 | |c| compatibility_fully_decomposed(c).or_else(|| canonical_fully_decomposed(c)); | |
8faf50e0 XL |
37 | decompose(c, decompose_char, emit_char) |
38 | } | |
abe05a73 | 39 | |
6a06907d XL |
40 | /// Compute standard-variation decomposition for character. |
41 | /// | |
42 | /// [Standardized Variation Sequences] are used instead of the standard canonical | |
43 | /// decompositions, notably for CJK codepoints with singleton canonical decompositions, | |
44 | /// to avoid losing information. See the | |
45 | /// [Unicode Variation Sequence FAQ](http://unicode.org/faq/vs.html) and the | |
46 | /// "Other Enhancements" section of the | |
47 | /// [Unicode 6.3 Release Summary](https://www.unicode.org/versions/Unicode6.3.0/#Summary) | |
48 | /// for more information. | |
49 | #[inline] | |
50 | pub fn decompose_cjk_compat_variants<F>(c: char, mut emit_char: F) | |
51 | where | |
52 | F: FnMut(char), | |
53 | { | |
54 | // 7-bit ASCII never decomposes | |
55 | if c <= '\x7f' { | |
56 | emit_char(c); | |
57 | return; | |
58 | } | |
59 | ||
60 | // Don't perform decomposition for Hangul | |
61 | ||
62 | if let Some(decomposed) = cjk_compat_variants_fully_decomposed(c) { | |
63 | for &d in decomposed { | |
64 | emit_char(d); | |
65 | } | |
66 | return; | |
67 | } | |
68 | ||
69 | // Finally bottom out. | |
70 | emit_char(c); | |
71 | } | |
72 | ||
8faf50e0 XL |
73 | #[inline] |
74 | fn decompose<D, F>(c: char, decompose_char: D, mut emit_char: F) | |
f035d41b XL |
75 | where |
76 | D: Fn(char) -> Option<&'static [char]>, | |
77 | F: FnMut(char), | |
8faf50e0 | 78 | { |
abe05a73 | 79 | // 7-bit ASCII never decomposes |
8faf50e0 XL |
80 | if c <= '\x7f' { |
81 | emit_char(c); | |
abe05a73 XL |
82 | return; |
83 | } | |
84 | ||
8faf50e0 XL |
85 | // Perform decomposition for Hangul |
86 | if is_hangul_syllable(c) { | |
87 | decompose_hangul(c, emit_char); | |
88 | return; | |
abe05a73 XL |
89 | } |
90 | ||
8faf50e0 XL |
91 | if let Some(decomposed) = decompose_char(c) { |
92 | for &d in decomposed { | |
93 | emit_char(d); | |
abe05a73 | 94 | } |
8faf50e0 | 95 | return; |
abe05a73 XL |
96 | } |
97 | ||
98 | // Finally bottom out. | |
8faf50e0 | 99 | emit_char(c); |
abe05a73 XL |
100 | } |
101 | ||
102 | /// Compose two characters into a single character, if possible. | |
103 | /// See [Unicode Standard Annex #15](http://www.unicode.org/reports/tr15/) | |
104 | /// for more information. | |
105 | pub fn compose(a: char, b: char) -> Option<char> { | |
dfeec247 | 106 | compose_hangul(a, b).or_else(|| composition_table(a, b)) |
abe05a73 XL |
107 | } |
108 | ||
109 | // Constants from Unicode 9.0.0 Section 3.12 Conjoining Jamo Behavior | |
110 | // http://www.unicode.org/versions/Unicode9.0.0/ch03.pdf#M9.32468.Heading.310.Combining.Jamo.Behavior | |
111 | const S_BASE: u32 = 0xAC00; | |
112 | const L_BASE: u32 = 0x1100; | |
113 | const V_BASE: u32 = 0x1161; | |
114 | const T_BASE: u32 = 0x11A7; | |
115 | const L_COUNT: u32 = 19; | |
116 | const V_COUNT: u32 = 21; | |
117 | const T_COUNT: u32 = 28; | |
f035d41b XL |
118 | const N_COUNT: u32 = V_COUNT * T_COUNT; |
119 | const S_COUNT: u32 = L_COUNT * N_COUNT; | |
abe05a73 | 120 | |
8faf50e0 XL |
121 | const S_LAST: u32 = S_BASE + S_COUNT - 1; |
122 | const L_LAST: u32 = L_BASE + L_COUNT - 1; | |
123 | const V_LAST: u32 = V_BASE + V_COUNT - 1; | |
124 | const T_LAST: u32 = T_BASE + T_COUNT - 1; | |
125 | ||
126 | // Composition only occurs for `TPart`s in `U+11A8 ... U+11C2`, | |
127 | // i.e. `T_BASE + 1 ... T_LAST`. | |
128 | const T_FIRST: u32 = T_BASE + 1; | |
129 | ||
130 | pub(crate) fn is_hangul_syllable(c: char) -> bool { | |
131 | (c as u32) >= S_BASE && (c as u32) < (S_BASE + S_COUNT) | |
132 | } | |
133 | ||
abe05a73 XL |
134 | // Decompose a precomposed Hangul syllable |
135 | #[allow(unsafe_code)] | |
136 | #[inline(always)] | |
f035d41b XL |
137 | fn decompose_hangul<F>(s: char, mut emit_char: F) |
138 | where | |
139 | F: FnMut(char), | |
140 | { | |
8faf50e0 XL |
141 | let s_index = s as u32 - S_BASE; |
142 | let l_index = s_index / N_COUNT; | |
abe05a73 | 143 | unsafe { |
8faf50e0 | 144 | emit_char(char::from_u32_unchecked(L_BASE + l_index)); |
abe05a73 | 145 | |
8faf50e0 XL |
146 | let v_index = (s_index % N_COUNT) / T_COUNT; |
147 | emit_char(char::from_u32_unchecked(V_BASE + v_index)); | |
abe05a73 | 148 | |
8faf50e0 XL |
149 | let t_index = s_index % T_COUNT; |
150 | if t_index > 0 { | |
151 | emit_char(char::from_u32_unchecked(T_BASE + t_index)); | |
abe05a73 XL |
152 | } |
153 | } | |
154 | } | |
155 | ||
8faf50e0 XL |
156 | #[inline] |
157 | pub(crate) fn hangul_decomposition_length(s: char) -> usize { | |
158 | let si = s as u32 - S_BASE; | |
159 | let ti = si % T_COUNT; | |
f035d41b XL |
160 | if ti > 0 { |
161 | 3 | |
162 | } else { | |
163 | 2 | |
164 | } | |
8faf50e0 XL |
165 | } |
166 | ||
abe05a73 XL |
167 | // Compose a pair of Hangul Jamo |
168 | #[allow(unsafe_code)] | |
169 | #[inline(always)] | |
dfeec247 | 170 | #[allow(ellipsis_inclusive_range_patterns)] |
abe05a73 | 171 | fn compose_hangul(a: char, b: char) -> Option<char> { |
8faf50e0 XL |
172 | let (a, b) = (a as u32, b as u32); |
173 | match (a, b) { | |
174 | // Compose a leading consonant and a vowel together into an LV_Syllable | |
f035d41b | 175 | (L_BASE...L_LAST, V_BASE...V_LAST) => { |
8faf50e0 XL |
176 | let l_index = a - L_BASE; |
177 | let v_index = b - V_BASE; | |
178 | let lv_index = l_index * N_COUNT + v_index * T_COUNT; | |
179 | let s = S_BASE + lv_index; | |
f035d41b XL |
180 | Some(unsafe { char::from_u32_unchecked(s) }) |
181 | } | |
8faf50e0 | 182 | // Compose an LV_Syllable and a trailing consonant into an LVT_Syllable |
f035d41b XL |
183 | (S_BASE...S_LAST, T_FIRST...T_LAST) if (a - S_BASE) % T_COUNT == 0 => { |
184 | Some(unsafe { char::from_u32_unchecked(a + (b - T_BASE)) }) | |
185 | } | |
8faf50e0 | 186 | _ => None, |
abe05a73 | 187 | } |
8faf50e0 XL |
188 | } |
189 | ||
190 | #[cfg(test)] | |
191 | mod tests { | |
192 | use super::compose_hangul; | |
193 | ||
194 | // Regression test from a bugfix where we were composing an LV_Syllable with | |
195 | // T_BASE directly. (We should only compose an LV_Syllable with a character | |
196 | // in the range `T_BASE + 1 ... T_LAST`.) | |
197 | #[test] | |
198 | fn test_hangul_composition() { | |
199 | assert_eq!(compose_hangul('\u{c8e0}', '\u{11a7}'), None); | |
abe05a73 | 200 | } |
abe05a73 | 201 | } |