xref: /plan9/sys/src/cmd/aux/antiword/utf8.c (revision 25b329d522281a8cdd35da0dcc08c3fc621059a9)
1 /*
2  * utf8.c
3  * Copyright (C) 2001-2004 A.J. van Os; Released under GPL
4  *
5  *====================================================================
6  * This part of the software is based on:
7  * An implementation of wcwidth() as defined in
8  * "The Single UNIX Specification, Version 2, The Open Group, 1997"
9  * <http://www.UNIX-systems.org/online.html>
10  * Markus Kuhn -- 2001-01-12 -- public domain
11  *====================================================================
12  * The credit should go to him, but all the bugs are mine.
13  */
14 
15 #include <stdlib.h>
16 #include <string.h>
17 #include "antiword.h"
18 
19 struct interval {
20 	USHORT	first;
21 	USHORT	last;
22 };
23 /* Sorted list of non-overlapping intervals of non-spacing characters */
24 static const struct interval combining[] = {
25 	{ 0x0300, 0x034E }, { 0x0360, 0x0362 }, { 0x0483, 0x0486 },
26 	{ 0x0488, 0x0489 }, { 0x0591, 0x05A1 }, { 0x05A3, 0x05B9 },
27 	{ 0x05BB, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
28 	{ 0x05C4, 0x05C4 }, { 0x064B, 0x0655 }, { 0x0670, 0x0670 },
29 	{ 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
30 	{ 0x070F, 0x070F }, { 0x0711, 0x0711 }, { 0x0730, 0x074A },
31 	{ 0x07A6, 0x07B0 }, { 0x0901, 0x0902 }, { 0x093C, 0x093C },
32 	{ 0x0941, 0x0948 }, { 0x094D, 0x094D }, { 0x0951, 0x0954 },
33 	{ 0x0962, 0x0963 }, { 0x0981, 0x0981 }, { 0x09BC, 0x09BC },
34 	{ 0x09C1, 0x09C4 }, { 0x09CD, 0x09CD }, { 0x09E2, 0x09E3 },
35 	{ 0x0A02, 0x0A02 }, { 0x0A3C, 0x0A3C }, { 0x0A41, 0x0A42 },
36 	{ 0x0A47, 0x0A48 }, { 0x0A4B, 0x0A4D }, { 0x0A70, 0x0A71 },
37 	{ 0x0A81, 0x0A82 }, { 0x0ABC, 0x0ABC }, { 0x0AC1, 0x0AC5 },
38 	{ 0x0AC7, 0x0AC8 }, { 0x0ACD, 0x0ACD }, { 0x0B01, 0x0B01 },
39 	{ 0x0B3C, 0x0B3C }, { 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 },
40 	{ 0x0B4D, 0x0B4D }, { 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 },
41 	{ 0x0BC0, 0x0BC0 }, { 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 },
42 	{ 0x0C46, 0x0C48 }, { 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 },
43 	{ 0x0CBF, 0x0CBF }, { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD },
44 	{ 0x0D41, 0x0D43 }, { 0x0D4D, 0x0D4D }, { 0x0DCA, 0x0DCA },
45 	{ 0x0DD2, 0x0DD4 }, { 0x0DD6, 0x0DD6 }, { 0x0E31, 0x0E31 },
46 	{ 0x0E34, 0x0E3A }, { 0x0E47, 0x0E4E }, { 0x0EB1, 0x0EB1 },
47 	{ 0x0EB4, 0x0EB9 }, { 0x0EBB, 0x0EBC }, { 0x0EC8, 0x0ECD },
48 	{ 0x0F18, 0x0F19 }, { 0x0F35, 0x0F35 }, { 0x0F37, 0x0F37 },
49 	{ 0x0F39, 0x0F39 }, { 0x0F71, 0x0F7E }, { 0x0F80, 0x0F84 },
50 	{ 0x0F86, 0x0F87 }, { 0x0F90, 0x0F97 }, { 0x0F99, 0x0FBC },
51 	{ 0x0FC6, 0x0FC6 }, { 0x102D, 0x1030 }, { 0x1032, 0x1032 },
52 	{ 0x1036, 0x1037 }, { 0x1039, 0x1039 }, { 0x1058, 0x1059 },
53 	{ 0x1160, 0x11FF }, { 0x17B7, 0x17BD }, { 0x17C6, 0x17C6 },
54 	{ 0x17C9, 0x17D3 }, { 0x180B, 0x180E }, { 0x18A9, 0x18A9 },
55 	{ 0x200B, 0x200F }, { 0x202A, 0x202E }, { 0x206A, 0x206F },
56 	{ 0x20D0, 0x20E3 }, { 0x302A, 0x302F }, { 0x3099, 0x309A },
57 	{ 0xFB1E, 0xFB1E }, { 0xFE20, 0xFE23 }, { 0xFEFF, 0xFEFF },
58 	{ 0xFFF9, 0xFFFB }
59 };
60 
61 /* Auxiliary function for binary search in interval table */
62 static BOOL
bIsZeroWidthChar(ULONG ucs)63 bIsZeroWidthChar(ULONG ucs)
64 {
65 	int low = 0;
66 	int high = elementsof(combining) - 1;
67 	int mid;
68 
69 	if (ucs < (ULONG)combining[low].first ||
70 	    ucs > (ULONG)combining[high].last) {
71 		return FALSE;
72 	}
73 
74 	while (high >= low) {
75 		mid = (low + high) / 2;
76 		if (ucs > (ULONG)combining[mid].last) {
77 			low = mid + 1;
78 		} else if (ucs < (ULONG)combining[mid].first) {
79 			high = mid - 1;
80 		} else {
81 			return TRUE;
82 		}
83 	}
84 	return FALSE;
85 } /* end of bIsZeroWidthChar */
86 
87 /* The following functions define the column width of an ISO 10646
88  * character as follows:
89  *
90  *    - The null character (U+0000) has a column width of 0.
91  *
92  *    - Other C0/C1 control characters and DEL will lead to a return
93  *      value of -1.
94  *
95  *    - Non-spacing and enclosing combining characters (general
96  *      category code Mn or Me in the Unicode database) have a
97  *      column width of 0.
98  *
99  *    - Other format characters (general category code Cf in the Unicode
100  *      database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
101  *
102  *    - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
103  *      have a column width of 0.
104  *
105  *    - Spacing characters in the East Asian Wide (W) or East Asian
106  *      FullWidth (F) category as defined in Unicode Technical
107  *      Report #11 have a column width of 2.
108  *
109  *    - All remaining characters (including all printable
110  *      ISO 8859-1 and WGL4 characters, Unicode control characters,
111  *      etc.) have a column width of 1.
112  *
113  * This implementation assumes that all characters are encoded
114  * in ISO 10646.
115  *
116  * This function is not named wcwidth() to prevent name clashes
117  */
118 static int
iWcWidth(ULONG ucs)119 iWcWidth(ULONG ucs)
120 {
121 	/* Test for 8-bit control characters */
122 	if (ucs == 0) {
123 		return 0;
124 	}
125 	if (ucs < 0x20 || (ucs >= 0x7f && ucs < 0xa0)) {
126 		NO_DBG_HEX(ucs);
127 		return -1;
128 	}
129 
130 	/* Binary search in table of non-spacing characters */
131 	if (bIsZeroWidthChar(ucs)) {
132 		return 0;
133 	}
134 
135 	/* Ucs is not a combining or C0/C1 control character */
136 
137 	return 1 +
138 	(ucs >= 0x1100 &&
139 	 (ucs <= 0x115f ||                    /* Hangul Jamo init. consonants */
140 	  (ucs >= 0x2e80 && ucs <= 0xa4cf && (ucs & ~0x0011) != 0x300a &&
141 	   ucs != 0x303f) ||                  /* CJK ... Yi */
142 	  (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
143 	  (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility Ideographs */
144 	  (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
145 	  (ucs >= 0xff00 && ucs <= 0xff5f) || /* Fullwidth Forms */
146 	  (ucs >= 0xffe0 && ucs <= 0xffe6) ||
147 	  (ucs >= 0x20000 && ucs <= 0x2ffff)));
148 } /* end of iWcWidth */
149 
150 /*
151  * utf8_to_ucs - convert from UTF-8 to UCS
152  *
153  * Returns the UCS character,
154  * Fills in the number of bytes in the UTF-8 character
155  */
156 static ULONG
utf8_to_ucs(const char * p,int iStrLen,int * piUtfLen)157 utf8_to_ucs(const char *p, int iStrLen, int *piUtfLen)
158 {
159 	ULONG	ulUcs;
160 	int	iIndex, iCharLen;
161 
162 	fail(p == NULL || piUtfLen == NULL);
163 	fail(iStrLen < 1);
164 
165 	ulUcs = (ULONG)(UCHAR)p[0];
166 
167 	if (ulUcs < 0x80) {
168 		*piUtfLen = 1;
169 		return ulUcs;
170 	}
171 
172 	if (ulUcs < 0xe0){
173 		iCharLen = 2;
174 		ulUcs &= 0x1f;
175 	} else if (ulUcs < 0xf0){
176 		iCharLen = 3;
177 		ulUcs &= 0x0f;
178 	} else if (ulUcs < 0xf8){
179 		iCharLen = 4;
180 		ulUcs &= 0x07;
181 	} else if (ulUcs < 0xfc){
182 		iCharLen = 5;
183 		ulUcs &= 0x03;
184 	} else {
185 		iCharLen = 6;
186 		ulUcs &= 0x01;
187 	}
188 	for (iIndex = 1; iIndex < iCharLen; iIndex++) {
189 		ulUcs <<= 6;
190 		if (iIndex < iStrLen) {
191 			ulUcs |= (ULONG)(UCHAR)p[iIndex] & 0x3f;
192 		}
193 	}
194 	*piUtfLen = iCharLen;
195 	return ulUcs;
196 } /* end of utf8_to_ucs */
197 
198 /*
199  * utf8_strwidth - compute the string width of an UTF-8 string
200  *
201  * Returns the string width in columns
202  */
203 long
utf8_strwidth(const char * pcString,size_t tNumchars)204 utf8_strwidth(const char *pcString, size_t tNumchars)
205 {
206 	ULONG	ulUcs;
207 	long	lTotal;
208 	int	iToGo, iWidth, iUtflen;
209 
210 	fail(pcString == NULL || tNumchars > (size_t)INT_MAX);
211 
212 	lTotal = 0;
213 	iToGo = (int)tNumchars;
214 
215 	while (iToGo > 0 && *pcString != '\0') {
216 		ulUcs = utf8_to_ucs(pcString, iToGo, &iUtflen);
217 		iWidth = iWcWidth(ulUcs);
218 		if (iWidth > 0) {
219 			lTotal += iWidth;
220 		}
221 		pcString += iUtflen;
222 		iToGo -= iUtflen;
223 	}
224 	NO_DBG_DEC(lTotal);
225 	return lTotal;
226 } /* end of utf8_strwidth */
227 
228 /*
229  * utf8_chrlength - get the number of bytes in an UTF-8 character
230  *
231  * Returns the number of bytes
232  */
233 int
utf8_chrlength(const char * p)234 utf8_chrlength(const char *p)
235 {
236 	int	iUtflen;
237 
238 	fail(p == NULL);
239 
240 	iUtflen = -1;		/* Just to make sure */
241 	(void)utf8_to_ucs(p, INT_MAX, &iUtflen);
242 	NO_DBG_DEC(iUtflen);
243 	return iUtflen;
244 } /* end of utf8_chrlength */
245 
246 /*
247  * is_locale_utf8 - return TRUE if the locale is UTF-8
248  */
249 BOOL
is_locale_utf8(void)250 is_locale_utf8(void)
251 {
252 	char	szCodeset[20];
253 
254 	szCodeset[0] = '\0';
255 	if (!bGetNormalizedCodeset(szCodeset, sizeof(szCodeset), NULL)) {
256 		return FALSE;
257 	}
258 	DBG_MSG(szCodeset);
259 	return STREQ(szCodeset, "utf8");
260 } /* end of is_locale_utf8 */
261