8000 Merge pull request #10 from unicode-rs/9 · mgeisler/unicode-normalization@b3a331c · GitHub
[go: up one dir, main page]
Skip to content

Commit b3a331c

Browse files
SimonSapinSimonSapin
authored
Merge pull request unicode-rs#10 from unicode-rs/9
Update to Unicode 9.0.0
2 parents cd9a45e + 7299191 commit b3a331c

File tree

4 files changed

+605
-207
lines changed

4 files changed

+605
-207
lines changed

‎Cargo.toml

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
[package]
22

33
name = "unicode-normalization"
4-
version = "0.1.2"
4+
version = "0.1.3"
55
authors = ["kwantam <kwantam@gmail.com>"]
66

77
homepage = "https://github.com/unicode-rs/unicode-normalization"

‎src/normalize.rs

Lines changed: 7 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -102,7 +102,8 @@ pub fn compose(a: char, b: char) -> Option<char> {
102102
})
103103
}
104104

105-
// Constants from Unicode 7.0.0 Section 3.12 Conjoining Jamo Behavior
105+
// Constants from Unicode 9.0.0 Section 3.12 Conjoining Jamo Behavior
106+
// http://www.unicode.org/versions/Unicode9.0.0/ch03.pdf#M9.32468.Heading.310.Combining.Jamo.Behavior
106107
const S_BASE: u32 = 0xAC00;
107108
const L_BASE: u32 = 0x1100;
108109
const V_BASE: u32 = 0x1161;
@@ -145,12 +146,15 @@ fn compose_hangul(a: char, b: char) -> Option<char> {
145146
let l = a as u32;
146147
let v = b as u32;
147148
// Compose an LPart and a VPart
148-
if L_BASE <= l && l < (L_BASE + L_COUNT) && V_BASE <= v && v < (V_BASE + V_COUNT) {
149+
if L_BASE <= l && l < (L_BASE + L_COUNT) // l should be an L choseong jamo
150+
&& V_BASE <= v && v < (V_BASE + V_COUNT) { // v should be a V jungseong jamo
149151
let r = S_BASE + (l - L_BASE) * N_COUNT + (v - V_BASE) * T_COUNT;
150152
return unsafe { Some(transmute(r)) };
151153
}
152154
// Compose an LVPart and a TPart
153-
if S_BASE <= l && l <= (S_BASE+S_COUNT-T_COUNT) && T_BASE <= v && v < (T_BASE+T_COUNT) {
155+
if S_BASE <= l && l <= (S_BASE+S_COUNT-T_COUNT) // l should be a syllable block
156+
&& T_BASE <= v && v < (T_BASE+T_COUNT) // v should be a T jongseong jamo
157+
&& (l - S_BASE) % T_COUNT == 0 { // l should be an LV syllable block (not LVT)
154158
let r = l + (v - T_BASE);
155159
return unsafe { Some(transmute(r)) };
156160
}

0 commit comments

Comments
 (0)
0