1 /* decomp.c - Character decomposition.
3 * Copyright (C) 1999, 2000 Tom Tromey
4 * Copyright 2000 Red Hat, Inc.
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.
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.
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.
27 #include "gunidecomp.h"
29 #include "gunicodeprivate.h"
32 #define CC_PART1(Page, Char) \
33 ((combining_class_table_part1[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
34 ? (combining_class_table_part1[Page] - G_UNICODE_MAX_TABLE_INDEX) \
35 : (cclass_data[combining_class_table_part1[Page]][Char]))
37 #define CC_PART2(Page, Char) \
38 ((combining_class_table_part2[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
39 ? (combining_class_table_part2[Page] - G_UNICODE_MAX_TABLE_INDEX) \
40 : (cclass_data[combining_class_table_part2[Page]][Char]))
42 #define COMBINING_CLASS(Char) \
43 (((Char) <= G_UNICODE_LAST_CHAR_PART1) \
44 ? CC_PART1 ((Char) >> 8, (Char) & 0xff) \
45 : (((Char) >= 0xe0000 && (Char) <= G_UNICODE_LAST_CHAR) \
46 ? CC_PART2 (((Char) - 0xe0000) >> 8, (Char) & 0xff) \
50 _g_unichar_combining_class (gunichar uc)
52 return COMBINING_CLASS (uc);
56 * g_unicode_canonical_ordering:
57 * @string: a UCS-4 encoded string.
58 * @len: the maximum length of @string to use.
60 * Computes the canonical ordering of a string in-place.
61 * This rearranges decomposed characters in the string
62 * according to their combining classes. See the Unicode
63 * manual for more information.
66 g_unicode_canonical_ordering (gunichar *string,
76 last = COMBINING_CLASS (string[0]);
77 for (i = 0; i < len - 1; ++i)
79 int next = COMBINING_CLASS (string[i + 1]);
80 if (next != 0 && last > next)
83 /* Percolate item leftward through string. */
84 for (j = i + 1; j > 0; --j)
87 if (COMBINING_CLASS (string[j - 1]) <= next)
90 string[j] = string[j - 1];
94 /* We're re-entering the loop looking at the old
103 /* returns a pointer to a null-terminated UTF-8 string */
105 find_decomposition (gunichar ch,
109 int end = G_N_ELEMENTS (decomp_table);
111 if (ch >= decomp_table[start].ch &&
112 ch <= decomp_table[end - 1].ch)
116 int half = (start + end) / 2;
117 if (ch == decomp_table[half].ch)
123 offset = decomp_table[half].compat_offset;
124 if (offset == G_UNICODE_NOT_PRESENT_OFFSET)
125 offset = decomp_table[half].canon_offset;
129 offset = decomp_table[half].canon_offset;
130 if (offset == G_UNICODE_NOT_PRESENT_OFFSET)
134 return &(decomp_expansion_string[offset]);
136 else if (half == start)
138 else if (ch > decomp_table[half].ch)
149 * g_unicode_canonical_decomposition:
150 * @ch: a Unicode character.
151 * @result_len: location to store the length of the return value.
153 * Computes the canonical decomposition of a Unicode character.
155 * Return value: a newly allocated string of Unicode characters.
156 * @result_len is set to the resulting length of the string.
159 g_unicode_canonical_decomposition (gunichar ch,
162 const gchar *decomp = find_decomposition (ch, FALSE);
171 *result_len = g_utf8_strlen (decomp, -1);
172 r = g_malloc (*result_len * sizeof (gunichar));
174 for (p = decomp, i = 0; *p != '\0'; p = g_utf8_next_char (p), i++)
175 r[i] = g_utf8_get_char (p);
179 /* Not in our table. */
180 r = g_malloc (sizeof (gunichar));
185 /* Supposedly following the Unicode 2.1.9 table means that the
186 decompositions come out in canonical order. I haven't tested
187 this, but we rely on it here. */
191 #define CI(Page, Char) \
192 ((compose_table[Page] >= G_UNICODE_MAX_TABLE_INDEX) \
193 ? (compose_table[Page] - G_UNICODE_MAX_TABLE_INDEX) \
194 : (compose_data[compose_table[Page]][Char]))
196 #define COMPOSE_INDEX(Char) \
197 (((Char) > (G_UNICODE_LAST_CHAR)) ? 0 : CI((Char) >> 8, (Char) & 0xff))
204 gushort index_a, index_b;
206 index_a = COMPOSE_INDEX(a);
208 if (index_a >= COMPOSE_FIRST_SINGLE_START && index_a < COMPOSE_SECOND_START)
210 if (b == compose_first_single[index_a - COMPOSE_FIRST_SINGLE_START][0])
212 *result = compose_first_single[index_a - COMPOSE_FIRST_SINGLE_START][1];
219 index_b = COMPOSE_INDEX(b);
221 if (index_b >= COMPOSE_SECOND_SINGLE_START)
223 if (a == compose_second_single[index_b - COMPOSE_SECOND_SINGLE_START][0])
225 *result = compose_second_single[index_b - COMPOSE_SECOND_SINGLE_START][1];
232 if (index_a >= COMPOSE_FIRST_START && index_a < COMPOSE_FIRST_SINGLE_START &&
233 index_b >= COMPOSE_SECOND_START && index_b < COMPOSE_SECOND_SINGLE_START)
235 gunichar res = compose_array[index_a - COMPOSE_FIRST_START][index_b - COMPOSE_SECOND_START];
248 _g_utf8_normalize_wc (const gchar *str,
256 gboolean do_compat = (mode == G_NORMALIZE_NFKC ||
257 mode == G_NORMALIZE_NFKD);
258 gboolean do_compose = (mode == G_NORMALIZE_NFC ||
259 mode == G_NORMALIZE_NFKC);
263 while ((max_len < 0 || p < str + max_len) && *p)
265 gunichar wc = g_utf8_get_char (p);
267 const gchar *decomp = find_decomposition (wc, do_compat);
270 n_wc += g_utf8_strlen (decomp, -1);
274 p = g_utf8_next_char (p);
277 wc_buffer = g_new (gunichar, n_wc + 1);
282 while ((max_len < 0 || p < str + max_len) && *p)
284 gunichar wc = g_utf8_get_char (p);
287 gsize old_n_wc = n_wc;
289 decomp = find_decomposition (wc, do_compat);
294 for (pd = decomp; *pd != '\0'; pd = g_utf8_next_char (pd))
295 wc_buffer[n_wc++] = g_utf8_get_char (pd);
298 wc_buffer[n_wc++] = wc;
302 cc = COMBINING_CLASS (wc_buffer[old_n_wc]);
306 g_unicode_canonical_ordering (wc_buffer + last_start, n_wc - last_start);
307 last_start = old_n_wc;
311 p = g_utf8_next_char (p);
316 g_unicode_canonical_ordering (wc_buffer + last_start, n_wc - last_start);
322 /* All decomposed and reordered */
324 if (do_compose && n_wc > 0)
330 for (i = 0; i < n_wc; i++)
332 int cc = COMBINING_CLASS (wc_buffer[i]);
335 (last_cc == 0 || last_cc != cc) &&
336 combine (wc_buffer[last_start], wc_buffer[i],
337 &wc_buffer[last_start]))
339 for (j = i + 1; j < n_wc; j++)
340 wc_buffer[j-1] = wc_buffer[j];
347 last_cc = COMBINING_CLASS (wc_buffer[i-1]);
366 * @str: a UTF-8 encoded string.
367 * @len: length of @str, in bytes, or -1 if @str is nul-terminated.
368 * @mode: the type of normalization to perform.
370 * Converts a string into canonical form, standardizing
371 * such issues as whether a character with an accent
372 * is represented as a base character and combining
373 * accent or as a single precomposed character. You
374 * should generally call g_utf8_normalize() before
375 * comparing two Unicode strings.
377 * The normalization mode %G_NORMALIZE_DEFAULT only
378 * standardizes differences that do not affect the
379 * text content, such as the above-mentioned accent
380 * representation. %G_NORMALIZE_ALL also standardizes
381 * the "compatibility" characters in Unicode, such
382 * as SUPERSCRIPT THREE to the standard forms
383 * (in this case DIGIT THREE). Formatting information
384 * may be lost but for most text operations such
385 * characters should be considered the same.
386 * For example, g_utf8_collate() normalizes
387 * with %G_NORMALIZE_ALL as its first step.
389 * %G_NORMALIZE_DEFAULT_COMPOSE and %G_NORMALIZE_ALL_COMPOSE
390 * are like %G_NORMALIZE_DEFAULT and %G_NORMALIZE_ALL,
391 * but returned a result with composed forms rather
392 * than a maximally decomposed form. This is often
393 * useful if you intend to convert the string to
394 * a legacy encoding or pass it to a system with
395 * less capable Unicode handling.
397 * Return value: a newly allocated string, that is the
398 * normalized form of @str.
401 g_utf8_normalize (const gchar *str,
405 gunichar *result_wc = _g_utf8_normalize_wc (str, len, mode);
408 result = g_ucs4_to_utf8 (result_wc, -1, NULL, NULL, NULL);