uc.c

   1 #include <ctype.h>
   2 #include <stdio.h>
   3 #include <stdlib.h>
   4 #include <string.h>
   5 #include "vi.h"
   6
   7 #define LEN(a)          (sizeof(a) / sizeof((a)[0]))
   8
   9 /* return the length of a utf-8 character */
  10 int uc_len(char *s)
  11 {
  12         int c = (unsigned char) s[0];
  13         if (~c & 0x80)
  14                 return c > 0;
  15         if (~c & 0x20)
  16                 return 2;
  17         if (~c & 0x10)
  18                 return 3;
  19         if (~c & 0x80)
  20                 return 4;
  21         if (~c & 0x40)
  22                 return 5;
  23         if (~c & 0x20)
  24                 return 6;
  25         return 1;
  26 }
  27
  28 /* the number of utf-8 characters in s */
  29 int uc_slen(char *s)
  30 {
  31         int n;
  32         for (n = 0; *s; n++)
  33                 s = uc_end(s) + 1;
  34         return n;
  35 }
  36
  37 /* the unicode codepoint of the given utf-8 character */
  38 int uc_code(char *s)
  39 {
  40         int c = (unsigned char) s[0];
  41         int l;
  42         if (!(c & 0x80))
  43                 return c;
  44         if (!(c & 0x20))
  45                 return ((c & 0x1f) << 6) | (s[1] & 0x3f);
  46         if (!(c & 0x10))
  47                 return ((c & 0x0f) << 12) | ((s[1] & 0x3f) << 6) | (s[2] & 0x3f);
  48         l = uc_len(s);
  49         c = (0x3f >> --l) & (unsigned char) *s++;
  50         while (l--)
  51                 c = (c << 6) | ((unsigned char) *s++ & 0x3f);
  52         return c;
  53 }
  54
  55 /* find the beginning of the character at s[i] */
  56 char *uc_beg(char *beg, char *s)
  57 {
  58         while (s > beg && (((unsigned char) *s) & 0xc0) == 0x80)
  59                 s--;
  60         return s;
  61 }
  62
  63 /* find the end of the character at s[i] */
  64 char *uc_end(char *s)
  65 {
  66         if (!*s || !((unsigned char) *s & 0x80))
  67                 return s;
  68         if (((unsigned char) *s & 0xc0) == 0xc0)
  69                 s++;
  70         while (((unsigned char) *s & 0xc0) == 0x80)
  71                 s++;
  72         return s - 1;
  73 }
  74
  75 /* return a pointer to the character following s */
  76 char *uc_next(char *s)
  77 {
  78         s = uc_end(s);
  79         return *s ? s + 1 : s;
  80 }
  81
  82 /* return a pointer to the character preceding s */
  83 char *uc_prev(char *beg, char *s)
  84 {
  85         return s == beg ? beg : uc_beg(beg, s - 1);
  86 }
  87
  88 int uc_wid(char *s)
  89 {
  90         return 1;
  91 }
  92
  93 char *uc_lastline(char *s)
  94 {
  95         char *r = strrchr(s, '\n');
  96         return r ? r + 1 : s;
  97 }
  98
  99 /* allocate and return an array for the characters in s */
 100 char **uc_chop(char *s, int *n)
 101 {
 102         char **chrs;
 103         int i;
 104         *n = uc_slen(s);
 105         chrs = malloc((*n + 1) * sizeof(chrs[0]));
 106         for (i = 0; i < *n + 1; i++) {
 107                 chrs[i] = s;
 108                 s = uc_next(s);
 109         }
 110         return chrs;
 111 }
 112
 113 char *uc_chr(char *s, int off)
 114 {
 115         int i = 0;
 116         while (s && *s) {
 117                 if (i++ == off)
 118                         return s;
 119                 s = uc_next(s);
 120         }
 121         return s && (off < 0 || i == off) ? s : "";
 122 }
 123
 124 /* the number of characters between s and s + off */
 125 int uc_off(char *s, int off)
 126 {
 127         char *e = s + off;
 128         int i;
 129         for (i = 0; s < e && *s; i++)
 130                 s = uc_next(s);
 131         return i;
 132 }
 133
 134 char *uc_sub(char *s, int beg, int end)
 135 {
 136         char *sbeg = uc_chr(s, beg);
 137         char *send = uc_chr(s, end);
 138         int len = sbeg && send && sbeg <= send ? send - sbeg : 0;
 139         char *r = malloc(len + 1);
 140         memcpy(r, sbeg, len);
 141         r[len] = '\0';
 142         return r;
 143 }
 144
 145 char *uc_dup(char *s)
 146 {
 147         char *r = malloc(strlen(s) + 1);
 148         return r ? strcpy(r, s) : NULL;
 149 }
 150
 151 int uc_isspace(char *s)
 152 {
 153         int c = s ? (unsigned char) *s : 0;
 154         return c <= 0x7f && isspace(c);
 155 }
 156
 157 int uc_isprint(char *s)
 158 {
 159         int c = s ? (unsigned char) *s : 0;
 160         return c > 0x7f || isprint(c);
 161 }
 162
 163 int uc_isalpha(char *s)
 164 {
 165         int c = s ? (unsigned char) *s : 0;
 166         return c > 0x7f || isalpha(c);
 167 }
 168
 169 int uc_isdigit(char *s)
 170 {
 171         int c = s ? (unsigned char) *s : 0;
 172         return c <= 0x7f && isdigit(c);
 173 }
 174
 175 int uc_kind(char *c)
 176 {
 177         if (uc_isspace(c))
 178                 return 0;
 179         if (uc_isalpha(c) || uc_isdigit(c) || c[0] == '_')
 180                 return 1;
 181         return 2;
 182 }
 183
 184 #define UC_R2L(ch)      (((ch) & 0xff00) == 0x0600 || \
 185                         ((ch) & 0xfffc) == 0x200c || \
 186                         ((ch) & 0xff00) == 0xfb00 || \
 187                         ((ch) & 0xff00) == 0xfc00 || \
 188                         ((ch) & 0xff00) == 0xfe00)
 189
 190 /* sorted list of characters that can be shaped */
 191 static struct achar {
 192         unsigned c;             /* utf-8 code */
 193         unsigned s;             /* single form */
 194         unsigned i;             /* initial form */
 195         unsigned m;             /* medial form */
 196         unsigned f;             /* final form */
 197 } achars[] = {
 198         {0x0621, 0xfe80},                               /* hamza */
 199         {0x0622, 0xfe81, 0, 0, 0xfe82},                 /* alef madda */
 200         {0x0623, 0xfe83, 0, 0, 0xfe84},                 /* alef hamza above */
 201         {0x0624, 0xfe85, 0, 0, 0xfe86},                 /* waw hamza */
 202         {0x0625, 0xfe87, 0, 0, 0xfe88},                 /* alef hamza below */
 203         {0x0626, 0xfe89, 0xfe8b, 0xfe8c, 0xfe8a},       /* yeh hamza */
 204         {0x0627, 0xfe8d, 0, 0, 0xfe8e},                 /* alef */
 205         {0x0628, 0xfe8f, 0xfe91, 0xfe92, 0xfe90},       /* beh */
 206         {0x0629, 0xfe93, 0, 0, 0xfe94},                 /* teh marbuta */
 207         {0x062a, 0xfe95, 0xfe97, 0xfe98, 0xfe96},       /* teh */
 208         {0x062b, 0xfe99, 0xfe9b, 0xfe9c, 0xfe9a},       /* theh */
 209         {0x062c, 0xfe9d, 0xfe9f, 0xfea0, 0xfe9e},       /* jeem */
 210         {0x062d, 0xfea1, 0xfea3, 0xfea4, 0xfea2},       /* hah */
 211         {0x062e, 0xfea5, 0xfea7, 0xfea8, 0xfea6},       /* khah */
 212         {0x062f, 0xfea9, 0, 0, 0xfeaa},                 /* dal */
 213         {0x0630, 0xfeab, 0, 0, 0xfeac},                 /* thal */
 214         {0x0631, 0xfead, 0, 0, 0xfeae},                 /* reh */
 215         {0x0632, 0xfeaf, 0, 0, 0xfeb0},                 /* zain */
 216         {0x0633, 0xfeb1, 0xfeb3, 0xfeb4, 0xfeb2},       /* seen */
 217         {0x0634, 0xfeb5, 0xfeb7, 0xfeb8, 0xfeb6},       /* sheen */
 218         {0x0635, 0xfeb9, 0xfebb, 0xfebc, 0xfeba},       /* sad */
 219         {0x0636, 0xfebd, 0xfebf, 0xfec0, 0xfebe},       /* dad */
 220         {0x0637, 0xfec1, 0xfec3, 0xfec4, 0xfec2},       /* tah */
 221         {0x0638, 0xfec5, 0xfec7, 0xfec8, 0xfec6},       /* zah */
 222         {0x0639, 0xfec9, 0xfecb, 0xfecc, 0xfeca},       /* ain */
 223         {0x063a, 0xfecd, 0xfecf, 0xfed0, 0xfece},       /* ghain */
 224         {0x0640, 0x640, 0x640, 0x640},                  /* tatweel */
 225         {0x0641, 0xfed1, 0xfed3, 0xfed4, 0xfed2},       /* feh */
 226         {0x0642, 0xfed5, 0xfed7, 0xfed8, 0xfed6},       /* qaf */
 227         {0x0643, 0xfed9, 0xfedb, 0xfedc, 0xfeda},       /* kaf */
 228         {0x0644, 0xfedd, 0xfedf, 0xfee0, 0xfede},       /* lam */
 229         {0x0645, 0xfee1, 0xfee3, 0xfee4, 0xfee2},       /* meem */
 230         {0x0646, 0xfee5, 0xfee7, 0xfee8, 0xfee6},       /* noon */
 231         {0x0647, 0xfee9, 0xfeeb, 0xfeec, 0xfeea},       /* heh */
 232         {0x0648, 0xfeed, 0, 0, 0xfeee},                 /* waw */
 233         {0x0649, 0xfeef, 0, 0, 0xfef0},                 /* alef maksura */
 234         {0x064a, 0xfef1, 0xfef3, 0xfef4, 0xfef2},       /* yeh */
 235         {0x067e, 0xfb56, 0xfb58, 0xfb59, 0xfb57},       /* peh */
 236         {0x0686, 0xfb7a, 0xfb7c, 0xfb7d, 0xfb7b},       /* tcheh */
 237         {0x0698, 0xfb8a, 0, 0, 0xfb8b},                 /* jeh */
 238         {0x06a9, 0xfb8e, 0xfb90, 0xfb91, 0xfb8f},       /* fkaf */
 239         {0x06af, 0xfb92, 0xfb94, 0xfb95, 0xfb93},       /* gaf */
 240         {0x06cc, 0xfbfc, 0xfbfe, 0xfbff, 0xfbfd},       /* fyeh */
 241         {0x200c},                                       /* ZWNJ */
 242         {0x200d, 0, 0x200d, 0x200d},                    /* ZWJ */
 243 };
 244
 245 static struct achar *find_achar(int c)
 246 {
 247         int h, m, l;
 248         h = LEN(achars);
 249         l = 0;
 250         /* using binary search to find c */
 251         while (l < h) {
 252                 m = (h + l) >> 1;
 253                 if (achars[m].c == c)
 254                         return &achars[m];
 255                 if (c < achars[m].c)
 256                         h = m;
 257                 else
 258                         l = m + 1;
 259         }
 260         return NULL;
 261 }
 262
 263 static int can_join(int c1, int c2)
 264 {
 265         struct achar *a1 = find_achar(c1);
 266         struct achar *a2 = find_achar(c2);
 267         return a1 && a2 && (a1->i || a1->m) && (a2->f || a2->m);
 268 }
 269
 270 static int uc_cshape(int cur, int prev, int next)
 271 {
 272         int c = cur;
 273         int join_prev, join_next;
 274         struct achar *ac = find_achar(c);
 275         if (!ac)                /* ignore non-Arabic characters */
 276                 return c;
 277         join_prev = can_join(prev, c);
 278         join_next = can_join(c, next);
 279         if (join_prev && join_next)
 280                 c = ac->m;
 281         if (join_prev && !join_next)
 282                 c = ac->f;
 283         if (!join_prev && join_next)
 284                 c = ac->i;
 285         if (!join_prev && !join_next)
 286                 c = ac->c;      /* some fonts do not have a glyph for ac->s */
 287         return c ? c : cur;
 288 }
 289
 290 /*
 291  * return nonzero for Arabic combining characters
 292  *
 293  * The standard Arabic diacritics:
 294  * + 0x064b: fathatan
 295  * + 0x064c: dammatan
 296  * + 0x064d: kasratan
 297  * + 0x064e: fatha
 298  * + 0x064f: damma
 299  * + 0x0650: kasra
 300  * + 0x0651: shadda
 301  * + 0x0652: sukun
 302  * + 0x0653: madda above
 303  * + 0x0654: hamza above
 304  * + 0x0655: hamza below
 305  * + 0x0670: superscript alef
 306  */
 307 static int uc_comb(int c)
 308 {
 309         return (c >= 0x064b && c <= 0x0655) ||          /* the standard diacritics */
 310                 (c >= 0xfc5e && c <= 0xfc63) ||         /* shadda ligatures */
 311                 c == 0x0670;                            /* superscript alef */
 312 }
 313
 314 static void uc_cput(char *d, int c)
 315 {
 316         int l = 0;
 317         if (c > 0xffff) {
 318                 *d++ = 0xf0 | (c >> 18);
 319                 l = 3;
 320         } else if (c > 0x7ff) {
 321                 *d++ = 0xe0 | (c >> 12);
 322                 l = 2;
 323         } else if (c > 0x7f) {
 324                 *d++ = 0xc0 | (c >> 6);
 325                 l = 1;
 326         } else {
 327                 *d++ = c;
 328         }
 329         while (l--)
 330                 *d++ = 0x80 | ((c >> (l * 6)) & 0x3f);
 331         *d = '\0';
 332 }
 333
 334 /* shape the given arabic character; returns a static buffer */
 335 char *uc_shape(char *beg, char *s)
 336 {
 337         static char out[16];
 338         char *r;
 339         int prev = 0;
 340         int next = 0;
 341         int curr = uc_code(s);
 342         if (!curr || !UC_R2L(curr))
 343                 return NULL;
 344         r = s;
 345         while (r > beg) {
 346                 r = uc_beg(beg, r - 1);
 347                 if (!uc_comb(uc_code(r))) {
 348                         prev = uc_code(r);
 349                         break;
 350                 }
 351         }
 352         r = s;
 353         while (*r) {
 354                 r = uc_next(r);
 355                 if (!uc_comb(uc_code(r))) {
 356                         next = uc_code(r);
 357                         break;
 358                 }
 359         }
 360         uc_cput(out, uc_cshape(curr, prev, next));
 361         return out;
 362 }