1 /* Test of compatibility decomposition of UTF-16 strings.
2 Copyright (C) 2009-2014 Free Software Foundation, Inc.
3
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 3 of the License, or
7 (at your option) any later version.
8
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
13
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
16
17 /* Written by Bruno Haible <bruno@clisp.org>, 2009. */
18
19 #include <config.h>
20
21 #if GNULIB_TEST_UNINORM_U16_NORMALIZE
22
23 #include "uninorm.h"
24
25 #include <signal.h>
26 #include <stdlib.h>
27 #include <unistd.h>
28
29 #include "unistr.h"
30 #include "macros.h"
31
32 static int
check(const uint16_t * input,size_t input_length,const uint16_t * expected,size_t expected_length)33 check (const uint16_t *input, size_t input_length,
34 const uint16_t *expected, size_t expected_length)
35 {
36 size_t length;
37 uint16_t *result;
38
39 /* Test return conventions with resultbuf == NULL. */
40 result = u16_normalize (UNINORM_NFKD, input, input_length, NULL, &length);
41 if (!(result != NULL))
42 return 1;
43 if (!(length == expected_length))
44 return 2;
45 if (!(u16_cmp (result, expected, expected_length) == 0))
46 return 3;
47 free (result);
48
49 /* Test return conventions with resultbuf too small. */
50 if (expected_length > 0)
51 {
52 uint16_t *preallocated;
53
54 length = expected_length - 1;
55 preallocated = (uint16_t *) malloc (length * sizeof (uint16_t));
56 result = u16_normalize (UNINORM_NFKD, input, input_length, preallocated, &length);
57 if (!(result != NULL))
58 return 4;
59 if (!(result != preallocated))
60 return 5;
61 if (!(length == expected_length))
62 return 6;
63 if (!(u16_cmp (result, expected, expected_length) == 0))
64 return 7;
65 free (result);
66 free (preallocated);
67 }
68
69 /* Test return conventions with resultbuf large enough. */
70 {
71 uint16_t *preallocated;
72
73 length = expected_length;
74 preallocated = (uint16_t *) malloc (length * sizeof (uint16_t));
75 result = u16_normalize (UNINORM_NFKD, input, input_length, preallocated, &length);
76 if (!(result != NULL))
77 return 8;
78 if (!(preallocated == NULL || result == preallocated))
79 return 9;
80 if (!(length == expected_length))
81 return 10;
82 if (!(u16_cmp (result, expected, expected_length) == 0))
83 return 11;
84 free (preallocated);
85 }
86
87 return 0;
88 }
89
90 void
test_u16_nfkd(void)91 test_u16_nfkd (void)
92 {
93 { /* Empty string. */
94 ASSERT (check (NULL, 0, NULL, 0) == 0);
95 }
96 { /* SPACE */
97 static const uint16_t input[] = { 0x0020 };
98 ASSERT (check (input, SIZEOF (input), input, SIZEOF (input)) == 0);
99 }
100
101 { /* LATIN CAPITAL LETTER A WITH DIAERESIS */
102 static const uint16_t input[] = { 0x00C4 };
103 static const uint16_t expected[] = { 0x0041, 0x0308 };
104 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
105 }
106
107 { /* LATIN CAPITAL LETTER A WITH DIAERESIS AND MACRON */
108 static const uint16_t input[] = { 0x01DE };
109 static const uint16_t expected[] = { 0x0041, 0x0308, 0x0304 };
110 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
111 }
112
113 { /* GREEK DIALYTIKA AND PERISPOMENI */
114 static const uint16_t input[] = { 0x1FC1 };
115 static const uint16_t expected[] = { 0x0020, 0x0308, 0x0342 };
116 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
117 }
118
119 { /* SCRIPT SMALL L */
120 static const uint16_t input[] = { 0x2113 };
121 static const uint16_t expected[] = { 0x006C };
122 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
123 }
124
125 { /* NO-BREAK SPACE */
126 static const uint16_t input[] = { 0x00A0 };
127 static const uint16_t expected[] = { 0x0020 };
128 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
129 }
130
131 { /* ARABIC LETTER VEH INITIAL FORM */
132 static const uint16_t input[] = { 0xFB6C };
133 static const uint16_t expected[] = { 0x06A4 };
134 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
135 }
136
137 { /* ARABIC LETTER VEH MEDIAL FORM */
138 static const uint16_t input[] = { 0xFB6D };
139 static const uint16_t expected[] = { 0x06A4 };
140 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
141 }
142
143 { /* ARABIC LETTER VEH FINAL FORM */
144 static const uint16_t input[] = { 0xFB6B };
145 static const uint16_t expected[] = { 0x06A4 };
146 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
147 }
148
149 { /* ARABIC LETTER VEH ISOLATED FORM */
150 static const uint16_t input[] = { 0xFB6A };
151 static const uint16_t expected[] = { 0x06A4 };
152 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
153 }
154
155 { /* CIRCLED NUMBER FIFTEEN */
156 static const uint16_t input[] = { 0x246E };
157 static const uint16_t expected[] = { 0x0031, 0x0035 };
158 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
159 }
160
161 { /* TRADE MARK SIGN */
162 static const uint16_t input[] = { 0x2122 };
163 static const uint16_t expected[] = { 0x0054, 0x004D };
164 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
165 }
166
167 { /* LATIN SUBSCRIPT SMALL LETTER I */
168 static const uint16_t input[] = { 0x1D62 };
169 static const uint16_t expected[] = { 0x0069 };
170 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
171 }
172
173 { /* PRESENTATION FORM FOR VERTICAL LEFT PARENTHESIS */
174 static const uint16_t input[] = { 0xFE35 };
175 static const uint16_t expected[] = { 0x0028 };
176 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
177 }
178
179 { /* FULLWIDTH LATIN CAPITAL LETTER A */
180 static const uint16_t input[] = { 0xFF21 };
181 static const uint16_t expected[] = { 0x0041 };
182 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
183 }
184
185 { /* HALFWIDTH IDEOGRAPHIC COMMA */
186 static const uint16_t input[] = { 0xFF64 };
187 static const uint16_t expected[] = { 0x3001 };
188 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
189 }
190
191 { /* SMALL IDEOGRAPHIC COMMA */
192 static const uint16_t input[] = { 0xFE51 };
193 static const uint16_t expected[] = { 0x3001 };
194 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
195 }
196
197 { /* SQUARE MHZ */
198 static const uint16_t input[] = { 0x3392 };
199 static const uint16_t expected[] = { 0x004D, 0x0048, 0x007A };
200 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
201 }
202
203 { /* VULGAR FRACTION THREE EIGHTHS */
204 static const uint16_t input[] = { 0x215C };
205 static const uint16_t expected[] = { 0x0033, 0x2044, 0x0038 };
206 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
207 }
208
209 { /* MICRO SIGN */
210 static const uint16_t input[] = { 0x00B5 };
211 static const uint16_t expected[] = { 0x03BC };
212 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
213 }
214
215 { /* ARABIC LIGATURE SALLALLAHOU ALAYHE WASALLAM */
216 static const uint16_t input[] = { 0xFDFA };
217 static const uint16_t expected[] =
218 { 0x0635, 0x0644, 0x0649, 0x0020, 0x0627, 0x0644, 0x0644, 0x0647, 0x0020,
219 0x0639, 0x0644, 0x064A, 0x0647, 0x0020, 0x0648, 0x0633, 0x0644, 0x0645
220 };
221 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
222 }
223
224 { /* HANGUL SYLLABLE GEUL */
225 static const uint16_t input[] = { 0xAE00 };
226 static const uint16_t expected[] = { 0x1100, 0x1173, 0x11AF };
227 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
228 }
229
230 { /* HANGUL SYLLABLE GEU */
231 static const uint16_t input[] = { 0xADF8 };
232 static const uint16_t expected[] = { 0x1100, 0x1173 };
233 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
234 }
235
236 { /* "Grüß Gott. Здравствуйте! x=(-b±sqrt(b²-4ac))/(2a) 日本語,中文,한글" */
237 static const uint16_t input[] =
238 { 'G', 'r', 0x00FC, 0x00DF, ' ', 'G', 'o', 't', 't', '.', ' ',
239 0x0417, 0x0434, 0x0440, 0x0430, 0x0432, 0x0441, 0x0442, 0x0432, 0x0443,
240 0x0439, 0x0442, 0x0435, '!', ' ',
241 'x', '=', '(', '-', 'b', 0x00B1, 's', 'q', 'r', 't', '(', 'b', 0x00B2,
242 '-', '4', 'a', 'c', ')', ')', '/', '(', '2', 'a', ')', ' ', ' ',
243 0x65E5, 0x672C, 0x8A9E, ',', 0x4E2D, 0x6587, ',', 0xD55C, 0xAE00, '\n'
244 };
245 static const uint16_t expected[] =
246 { 'G', 'r', 0x0075, 0x0308, 0x00DF, ' ', 'G', 'o', 't', 't', '.', ' ',
247 0x0417, 0x0434, 0x0440, 0x0430, 0x0432, 0x0441, 0x0442, 0x0432, 0x0443,
248 0x0438, 0x0306, 0x0442, 0x0435, '!', ' ',
249 'x', '=', '(', '-', 'b', 0x00B1, 's', 'q', 'r', 't', '(', 'b', 0x0032,
250 '-', '4', 'a', 'c', ')', ')', '/', '(', '2', 'a', ')', ' ', ' ',
251 0x65E5, 0x672C, 0x8A9E, ',', 0x4E2D, 0x6587, ',',
252 0x1112, 0x1161, 0x11AB, 0x1100, 0x1173, 0x11AF, '\n'
253 };
254 ASSERT (check (input, SIZEOF (input), expected, SIZEOF (expected)) == 0);
255 }
256
257 #if HAVE_DECL_ALARM
258 /* Declare failure if test takes too long, by using default abort
259 caused by SIGALRM. */
260 signal (SIGALRM, SIG_DFL);
261 alarm (50);
262 #endif
263
264 /* Check that the sorting is not O(n²) but O(n log n). */
265 {
266 int pass;
267 for (pass = 0; pass < 3; pass++)
268 {
269 size_t repeat = 1;
270 size_t m = 100000;
271 uint16_t *input = (uint16_t *) malloc (2 * m * sizeof (uint16_t));
272 if (input != NULL)
273 {
274 uint16_t *expected = input + m;
275 size_t m1 = m / 2;
276 size_t m2 = (m - 1) / 2;
277 /* NB: m1 + m2 == m - 1. */
278 uint16_t *p;
279 size_t i;
280
281 input[0] = 0x0041;
282 p = input + 1;
283 switch (pass)
284 {
285 case 0:
286 for (i = 0; i < m1; i++)
287 *p++ = 0x0319;
288 for (i = 0; i < m2; i++)
289 *p++ = 0x0300;
290 break;
291
292 case 1:
293 for (i = 0; i < m2; i++)
294 *p++ = 0x0300;
295 for (i = 0; i < m1; i++)
296 *p++ = 0x0319;
297 break;
298
299 case 2:
300 for (i = 0; i < m2; i++)
301 {
302 *p++ = 0x0319;
303 *p++ = 0x0300;
304 }
305 for (; i < m1; i++)
306 *p++ = 0x0319;
307 break;
308
309 default:
310 abort ();
311 }
312
313 expected[0] = 0x0041;
314 p = expected + 1;
315 for (i = 0; i < m1; i++)
316 *p++ = 0x0319;
317 for (i = 0; i < m2; i++)
318 *p++ = 0x0300;
319
320 for (; repeat > 0; repeat--)
321 ASSERT (check (input, m, expected, m) == 0);
322
323 free (input);
324 }
325 }
326 }
327 }
328
329 #else
330
331 void
test_u16_nfkd(void)332 test_u16_nfkd (void)
333 {
334 }
335
336 #endif
337