• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* decomp.c - Character decomposition.
2  *
3  *  Copyright (C) 1999, 2000 Tom Tromey
4  *  Copyright 2000 Red Hat, Inc.
5  *
6  * The Gnome Library is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public License as
8  * published by the Free Software Foundation; either version 2 of the
9  * License, or (at your option) any later version.
10  *
11  * The Gnome Library is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with the Gnome Library; see the file COPYING.LIB.  If not,
18  * write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19  *   Boston, MA 02111-1307, USA.
20  */
21 
22 #include "config.h"
23 
24 #include <stdlib.h>
25 
26 #include "glib.h"
27 #include "gunidecomp.h"
28 #include "gunicomp.h"
29 #include "gunicodeprivate.h"
30 #include "galias.h"
31 
32 
33 #define CC_PART1(Page, Char) \
34   ((combining_class_table_part1[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
35    ? (combining_class_table_part1[Page] - G_UNICODE_MAX_TABLE_INDEX) \
36    : (cclass_data[combining_class_table_part1[Page]][Char]))
37 
38 #define CC_PART2(Page, Char) \
39   ((combining_class_table_part2[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
40    ? (combining_class_table_part2[Page] - G_UNICODE_MAX_TABLE_INDEX) \
41    : (cclass_data[combining_class_table_part2[Page]][Char]))
42 
43 #define COMBINING_CLASS(Char) \
44   (((Char) <= G_UNICODE_LAST_CHAR_PART1) \
45    ? CC_PART1 ((Char) >> 8, (Char) & 0xff) \
46    : (((Char) >= 0xe0000 && (Char) <= G_UNICODE_LAST_CHAR) \
47       ? CC_PART2 (((Char) - 0xe0000) >> 8, (Char) & 0xff) \
48       : 0))
49 
50 /**
51  * g_unichar_combining_class:
52  * @uc: a Unicode character
53  *
54  * Determines the canonical combining class of a Unicode character.
55  *
56  * Return value: the combining class of the character
57  *
58  * Since: 2.14
59  **/
60 gint
g_unichar_combining_class(gunichar uc)61 g_unichar_combining_class (gunichar uc)
62 {
63   return COMBINING_CLASS (uc);
64 }
65 
66 /* constants for hangul syllable [de]composition */
67 #define SBase 0xAC00
68 #define LBase 0x1100
69 #define VBase 0x1161
70 #define TBase 0x11A7
71 #define LCount 19
72 #define VCount 21
73 #define TCount 28
74 #define NCount (VCount * TCount)
75 #define SCount (LCount * NCount)
76 
77 /**
78  * g_unicode_canonical_ordering:
79  * @string: a UCS-4 encoded string.
80  * @len: the maximum length of @string to use.
81  *
82  * Computes the canonical ordering of a string in-place.
83  * This rearranges decomposed characters in the string
84  * according to their combining classes.  See the Unicode
85  * manual for more information.
86  **/
87 void
g_unicode_canonical_ordering(gunichar * string,gsize len)88 g_unicode_canonical_ordering (gunichar *string,
89 			      gsize     len)
90 {
91   gsize i;
92   int swap = 1;
93 
94   while (swap)
95     {
96       int last;
97       swap = 0;
98       last = COMBINING_CLASS (string[0]);
99       for (i = 0; i < len - 1; ++i)
100 	{
101 	  int next = COMBINING_CLASS (string[i + 1]);
102 	  if (next != 0 && last > next)
103 	    {
104 	      gsize j;
105 	      /* Percolate item leftward through string.  */
106 	      for (j = i + 1; j > 0; --j)
107 		{
108 		  gunichar t;
109 		  if (COMBINING_CLASS (string[j - 1]) <= next)
110 		    break;
111 		  t = string[j];
112 		  string[j] = string[j - 1];
113 		  string[j - 1] = t;
114 		  swap = 1;
115 		}
116 	      /* We're re-entering the loop looking at the old
117 		 character again.  */
118 	      next = last;
119 	    }
120 	  last = next;
121 	}
122     }
123 }
124 
125 /* http://www.unicode.org/unicode/reports/tr15/#Hangul
126  * r should be null or have sufficient space. Calling with r == NULL will
127  * only calculate the result_len; however, a buffer with space for three
128  * characters will always be big enough. */
129 static void
decompose_hangul(gunichar s,gunichar * r,gsize * result_len)130 decompose_hangul (gunichar s,
131                   gunichar *r,
132                   gsize *result_len)
133 {
134   gint SIndex = s - SBase;
135 
136   /* not a hangul syllable */
137   if (SIndex < 0 || SIndex >= SCount)
138     {
139       if (r)
140         r[0] = s;
141       *result_len = 1;
142     }
143   else
144     {
145       gunichar L = LBase + SIndex / NCount;
146       gunichar V = VBase + (SIndex % NCount) / TCount;
147       gunichar T = TBase + SIndex % TCount;
148 
149       if (r)
150         {
151           r[0] = L;
152           r[1] = V;
153         }
154 
155       if (T != TBase)
156         {
157           if (r)
158             r[2] = T;
159           *result_len = 3;
160         }
161       else
162         *result_len = 2;
163     }
164 }
165 
166 /* returns a pointer to a null-terminated UTF-8 string */
167 static const gchar *
find_decomposition(gunichar ch,gboolean compat)168 find_decomposition (gunichar ch,
169 		    gboolean compat)
170 {
171   int start = 0;
172   int end = G_N_ELEMENTS (decomp_table);
173 
174   if (ch >= decomp_table[start].ch &&
175       ch <= decomp_table[end - 1].ch)
176     {
177       while (TRUE)
178 	{
179 	  int half = (start + end) / 2;
180 	  if (ch == decomp_table[half].ch)
181 	    {
182 	      int offset;
183 
184 	      if (compat)
185 		{
186 		  offset = decomp_table[half].compat_offset;
187 		  if (offset == G_UNICODE_NOT_PRESENT_OFFSET)
188 		    offset = decomp_table[half].canon_offset;
189 		}
190 	      else
191 		{
192 		  offset = decomp_table[half].canon_offset;
193 		  if (offset == G_UNICODE_NOT_PRESENT_OFFSET)
194 		    return NULL;
195 		}
196 
197 	      return &(decomp_expansion_string[offset]);
198 	    }
199 	  else if (half == start)
200 	    break;
201 	  else if (ch > decomp_table[half].ch)
202 	    start = half;
203 	  else
204 	    end = half;
205 	}
206     }
207 
208   return NULL;
209 }
210 
211 /**
212  * g_unicode_canonical_decomposition:
213  * @ch: a Unicode character.
214  * @result_len: location to store the length of the return value.
215  *
216  * Computes the canonical decomposition of a Unicode character.
217  *
218  * Return value: a newly allocated string of Unicode characters.
219  *   @result_len is set to the resulting length of the string.
220  **/
221 gunichar *
g_unicode_canonical_decomposition(gunichar ch,gsize * result_len)222 g_unicode_canonical_decomposition (gunichar ch,
223 				   gsize   *result_len)
224 {
225   const gchar *decomp;
226   const gchar *p;
227   gunichar *r;
228 
229   /* Hangul syllable */
230   if (ch >= 0xac00 && ch <= 0xd7a3)
231     {
232       decompose_hangul (ch, NULL, result_len);
233       r = g_malloc (*result_len * sizeof (gunichar));
234       decompose_hangul (ch, r, result_len);
235     }
236   else if ((decomp = find_decomposition (ch, FALSE)) != NULL)
237     {
238       /* Found it.  */
239       int i;
240 
241       *result_len = g_utf8_strlen (decomp, -1);
242       r = g_malloc (*result_len * sizeof (gunichar));
243 
244       for (p = decomp, i = 0; *p != '\0'; p = g_utf8_next_char (p), i++)
245         r[i] = g_utf8_get_char (p);
246     }
247   else
248     {
249       /* Not in our table.  */
250       r = g_malloc (sizeof (gunichar));
251       *r = ch;
252       *result_len = 1;
253     }
254 
255   /* Supposedly following the Unicode 2.1.9 table means that the
256      decompositions come out in canonical order.  I haven't tested
257      this, but we rely on it here.  */
258   return r;
259 }
260 
261 /* L,V => LV and LV,T => LVT  */
262 static gboolean
combine_hangul(gunichar a,gunichar b,gunichar * result)263 combine_hangul (gunichar a,
264                 gunichar b,
265                 gunichar *result)
266 {
267   gint LIndex = a - LBase;
268   gint SIndex = a - SBase;
269 
270   gint VIndex = b - VBase;
271   gint TIndex = b - TBase;
272 
273   if (0 <= LIndex && LIndex < LCount
274       && 0 <= VIndex && VIndex < VCount)
275     {
276       *result = SBase + (LIndex * VCount + VIndex) * TCount;
277       return TRUE;
278     }
279   else if (0 <= SIndex && SIndex < SCount && (SIndex % TCount) == 0
280            && 0 < TIndex && TIndex < TCount)
281     {
282       *result = a + TIndex;
283       return TRUE;
284     }
285 
286   return FALSE;
287 }
288 
289 #define CI(Page, Char) \
290   ((compose_table[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
291    ? (compose_table[Page] - G_UNICODE_MAX_TABLE_INDEX) \
292    : (compose_data[compose_table[Page]][Char]))
293 
294 #define COMPOSE_INDEX(Char) \
295      (((Char >> 8) > (COMPOSE_TABLE_LAST)) ? 0 : CI((Char) >> 8, (Char) & 0xff))
296 
297 static gboolean
combine(gunichar a,gunichar b,gunichar * result)298 combine (gunichar  a,
299 	 gunichar  b,
300 	 gunichar *result)
301 {
302   gushort index_a, index_b;
303 
304   if (combine_hangul (a, b, result))
305     return TRUE;
306 
307   index_a = COMPOSE_INDEX(a);
308 
309   if (index_a >= COMPOSE_FIRST_SINGLE_START && index_a < COMPOSE_SECOND_START)
310     {
311       if (b == compose_first_single[index_a - COMPOSE_FIRST_SINGLE_START][0])
312 	{
313 	  *result = compose_first_single[index_a - COMPOSE_FIRST_SINGLE_START][1];
314 	  return TRUE;
315 	}
316       else
317         return FALSE;
318     }
319 
320   index_b = COMPOSE_INDEX(b);
321 
322   if (index_b >= COMPOSE_SECOND_SINGLE_START)
323     {
324       if (a == compose_second_single[index_b - COMPOSE_SECOND_SINGLE_START][0])
325 	{
326 	  *result = compose_second_single[index_b - COMPOSE_SECOND_SINGLE_START][1];
327 	  return TRUE;
328 	}
329       else
330         return FALSE;
331     }
332 
333   if (index_a >= COMPOSE_FIRST_START && index_a < COMPOSE_FIRST_SINGLE_START &&
334       index_b >= COMPOSE_SECOND_START && index_b < COMPOSE_SECOND_SINGLE_START)
335     {
336       gunichar res = compose_array[index_a - COMPOSE_FIRST_START][index_b - COMPOSE_SECOND_START];
337 
338       if (res)
339 	{
340 	  *result = res;
341 	  return TRUE;
342 	}
343     }
344 
345   return FALSE;
346 }
347 
348 gunichar *
_g_utf8_normalize_wc(const gchar * str,gssize max_len,GNormalizeMode mode)349 _g_utf8_normalize_wc (const gchar    *str,
350 		      gssize          max_len,
351 		      GNormalizeMode  mode)
352 {
353   gsize n_wc;
354   gunichar *wc_buffer;
355   const char *p;
356   gsize last_start;
357   gboolean do_compat = (mode == G_NORMALIZE_NFKC ||
358 			mode == G_NORMALIZE_NFKD);
359   gboolean do_compose = (mode == G_NORMALIZE_NFC ||
360 			 mode == G_NORMALIZE_NFKC);
361 
362   n_wc = 0;
363   p = str;
364   while ((max_len < 0 || p < str + max_len) && *p)
365     {
366       const gchar *decomp;
367       gunichar wc = g_utf8_get_char (p);
368 
369       if (wc >= 0xac00 && wc <= 0xd7a3)
370         {
371           gsize result_len;
372           decompose_hangul (wc, NULL, &result_len);
373           n_wc += result_len;
374         }
375       else
376         {
377           decomp = find_decomposition (wc, do_compat);
378 
379           if (decomp)
380             n_wc += g_utf8_strlen (decomp, -1);
381           else
382             n_wc++;
383         }
384 
385       p = g_utf8_next_char (p);
386     }
387 
388   wc_buffer = g_new (gunichar, n_wc + 1);
389 
390   last_start = 0;
391   n_wc = 0;
392   p = str;
393   while ((max_len < 0 || p < str + max_len) && *p)
394     {
395       gunichar wc = g_utf8_get_char (p);
396       const gchar *decomp;
397       int cc;
398       gsize old_n_wc = n_wc;
399 
400       if (wc >= 0xac00 && wc <= 0xd7a3)
401         {
402           gsize result_len;
403           decompose_hangul (wc, wc_buffer + n_wc, &result_len);
404           n_wc += result_len;
405         }
406       else
407         {
408           decomp = find_decomposition (wc, do_compat);
409 
410           if (decomp)
411             {
412               const char *pd;
413               for (pd = decomp; *pd != '\0'; pd = g_utf8_next_char (pd))
414                 wc_buffer[n_wc++] = g_utf8_get_char (pd);
415             }
416           else
417             wc_buffer[n_wc++] = wc;
418         }
419 
420       if (n_wc > 0)
421 	{
422 	  cc = COMBINING_CLASS (wc_buffer[old_n_wc]);
423 
424 	  if (cc == 0)
425 	    {
426 	      g_unicode_canonical_ordering (wc_buffer + last_start, n_wc - last_start);
427 	      last_start = old_n_wc;
428 	    }
429 	}
430 
431       p = g_utf8_next_char (p);
432     }
433 
434   if (n_wc > 0)
435     {
436       g_unicode_canonical_ordering (wc_buffer + last_start, n_wc - last_start);
437       last_start = n_wc;
438     }
439 
440   wc_buffer[n_wc] = 0;
441 
442   /* All decomposed and reordered */
443 
444   if (do_compose && n_wc > 0)
445     {
446       gsize i, j;
447       int last_cc = 0;
448       last_start = 0;
449 
450       for (i = 0; i < n_wc; i++)
451 	{
452 	  int cc = COMBINING_CLASS (wc_buffer[i]);
453 
454 	  if (i > 0 &&
455 	      (last_cc == 0 || last_cc < cc) &&
456 	      combine (wc_buffer[last_start], wc_buffer[i],
457 		       &wc_buffer[last_start]))
458 	    {
459 	      for (j = i + 1; j < n_wc; j++)
460 		wc_buffer[j-1] = wc_buffer[j];
461 	      n_wc--;
462 	      i--;
463 
464 	      if (i == last_start)
465 		last_cc = 0;
466 	      else
467 		last_cc = COMBINING_CLASS (wc_buffer[i-1]);
468 
469 	      continue;
470 	    }
471 
472 	  if (cc == 0)
473 	    last_start = i;
474 
475 	  last_cc = cc;
476 	}
477     }
478 
479   wc_buffer[n_wc] = 0;
480 
481   return wc_buffer;
482 }
483 
484 /**
485  * g_utf8_normalize:
486  * @str: a UTF-8 encoded string.
487  * @len: length of @str, in bytes, or -1 if @str is nul-terminated.
488  * @mode: the type of normalization to perform.
489  *
490  * Converts a string into canonical form, standardizing
491  * such issues as whether a character with an accent
492  * is represented as a base character and combining
493  * accent or as a single precomposed character. The
494  * string has to be valid UTF-8, otherwise %NULL is
495  * returned. You should generally call g_utf8_normalize()
496  * before comparing two Unicode strings.
497  *
498  * The normalization mode %G_NORMALIZE_DEFAULT only
499  * standardizes differences that do not affect the
500  * text content, such as the above-mentioned accent
501  * representation. %G_NORMALIZE_ALL also standardizes
502  * the "compatibility" characters in Unicode, such
503  * as SUPERSCRIPT THREE to the standard forms
504  * (in this case DIGIT THREE). Formatting information
505  * may be lost but for most text operations such
506  * characters should be considered the same.
507  *
508  * %G_NORMALIZE_DEFAULT_COMPOSE and %G_NORMALIZE_ALL_COMPOSE
509  * are like %G_NORMALIZE_DEFAULT and %G_NORMALIZE_ALL,
510  * but returned a result with composed forms rather
511  * than a maximally decomposed form. This is often
512  * useful if you intend to convert the string to
513  * a legacy encoding or pass it to a system with
514  * less capable Unicode handling.
515  *
516  * Return value: a newly allocated string, that is the
517  *   normalized form of @str, or %NULL if @str is not
518  *   valid UTF-8.
519  **/
520 gchar *
g_utf8_normalize(const gchar * str,gssize len,GNormalizeMode mode)521 g_utf8_normalize (const gchar    *str,
522 		  gssize          len,
523 		  GNormalizeMode  mode)
524 {
525   gunichar *result_wc = _g_utf8_normalize_wc (str, len, mode);
526   gchar *result;
527 
528   result = g_ucs4_to_utf8 (result_wc, -1, NULL, NULL, NULL);
529   g_free (result_wc);
530 
531   return result;
532 }
533 
534 #define __G_UNIDECOMP_C__
535 #include "galiasdef.c"
536