utf8.c: use a table for double_width
Refactor git_wcwidth() and replace the if-else-if chain. Use the table double_width which is scanned by the bisearch() function, which is already used to find combining code points. Signed-off-by: Torsten Bögershausen <tboegi@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Torsten Bögershausen committed
May 9, 2014 at 23:51 UTC
08460345b5a221c0076a48c6be875d64b78b6015
1 file changed
+18
-23
utf8.c
+18
-23
@@ -126,6 +126,19 @@ static int git_wcwidth(ucs_char_t ch)
126
{ 0x1D1AA, 0x1D1AD }, { 0xE0001, 0xE0001 },
127
{ 0xE0020, 0xE007F }, { 0xE0100, 0xE01EF }
128
};
129
+ static const struct interval double_width[] = {
130
+ { 0x1100, 0x115F },
131
+ { 0x2329, 0x232A },
132
+ { 0x2E80, 0x303E },
133
+ { 0x3040, 0xA4CF },
134
+ { 0xAC00, 0xD7A3 },
135
+ { 0xF900, 0xFAFF },
136
+ { 0xFE30, 0xFE6F },
137
+ { 0xFF00, 0xFF60 },
138
+ { 0xFFE0, 0xFFE6 },
139
+ { 0x20000, 0x2FFFD },
140
+ { 0x30000, 0x3FFFD }
141
+ };
142
143
/* test for 8-bit control characters */
144
if (ch == 0)
@@ -138,30 +151,12 @@ static int git_wcwidth(ucs_char_t ch)
151
/ sizeof(struct interval) - 1))
152
return 0;
153
141
- /*
142
- * If we arrive here, ch is neither a combining nor a C0/C1
143
- * control character.
144
- */
154
+ /* binary search in table of double width characters */
155
+ if (bisearch(ch, double_width, sizeof(double_width)
156
+ / sizeof(struct interval) - 1))
157
+ return 2;
158
146
- return 1 +
147
- (ch >= 0x1100 &&
148
- /* Hangul Jamo init. consonants */
149
- (ch <= 0x115f ||
150
- ch == 0x2329 || ch == 0x232a ||
151
- /* CJK ... Yi */
152
- (ch >= 0x2e80 && ch <= 0xa4cf &&
153
- ch != 0x303f) ||
154
- /* Hangul Syllables */
155
- (ch >= 0xac00 && ch <= 0xd7a3) ||
156
- /* CJK Compatibility Ideographs */
157
- (ch >= 0xf900 && ch <= 0xfaff) ||
158
- /* CJK Compatibility Forms */
159
- (ch >= 0xfe30 && ch <= 0xfe6f) ||
160
- /* Fullwidth Forms */
161
- (ch >= 0xff00 && ch <= 0xff60) ||
162
- (ch >= 0xffe0 && ch <= 0xffe6) ||
163
- (ch >= 0x20000 && ch <= 0x2fffd) ||
164
- (ch >= 0x30000 && ch <= 0x3fffd)));
159
+ return 1;
160
}
161
162
/*