1 | /* Test of conversion of multibyte character to wide character.
|
---|
2 | Copyright (C) 2008-2012 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 <[email protected]>, 2008. */
|
---|
18 |
|
---|
19 | #include <config.h>
|
---|
20 |
|
---|
21 | #include <wchar.h>
|
---|
22 |
|
---|
23 | #include "signature.h"
|
---|
24 | SIGNATURE_CHECK (mbrtowc, size_t, (wchar_t *, char const *, size_t,
|
---|
25 | mbstate_t *));
|
---|
26 |
|
---|
27 | #include <locale.h>
|
---|
28 | #include <stdio.h>
|
---|
29 | #include <string.h>
|
---|
30 |
|
---|
31 | #include "macros.h"
|
---|
32 |
|
---|
33 | int
|
---|
34 | main (int argc, char *argv[])
|
---|
35 | {
|
---|
36 | mbstate_t state;
|
---|
37 | wchar_t wc;
|
---|
38 | size_t ret;
|
---|
39 |
|
---|
40 | /* configure should already have checked that the locale is supported. */
|
---|
41 | if (setlocale (LC_ALL, "") == NULL)
|
---|
42 | return 1;
|
---|
43 |
|
---|
44 | /* Test zero-length input. */
|
---|
45 | {
|
---|
46 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
47 | wc = (wchar_t) 0xBADFACE;
|
---|
48 | ret = mbrtowc (&wc, "x", 0, &state);
|
---|
49 | /* gnulib's implementation returns (size_t)(-2).
|
---|
50 | The AIX 5.1 implementation returns (size_t)(-1).
|
---|
51 | glibc's implementation returns 0. */
|
---|
52 | ASSERT (ret == (size_t)(-2) || ret == (size_t)(-1) || ret == 0);
|
---|
53 | ASSERT (mbsinit (&state));
|
---|
54 | }
|
---|
55 |
|
---|
56 | /* Test NUL byte input. */
|
---|
57 | {
|
---|
58 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
59 | wc = (wchar_t) 0xBADFACE;
|
---|
60 | ret = mbrtowc (&wc, "", 1, &state);
|
---|
61 | ASSERT (ret == 0);
|
---|
62 | ASSERT (wc == 0);
|
---|
63 | ASSERT (mbsinit (&state));
|
---|
64 | ret = mbrtowc (NULL, "", 1, &state);
|
---|
65 | ASSERT (ret == 0);
|
---|
66 | ASSERT (mbsinit (&state));
|
---|
67 | }
|
---|
68 |
|
---|
69 | /* Test single-byte input. */
|
---|
70 | {
|
---|
71 | int c;
|
---|
72 | char buf[1];
|
---|
73 |
|
---|
74 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
75 | for (c = 0; c < 0x100; c++)
|
---|
76 | switch (c)
|
---|
77 | {
|
---|
78 | case '\t': case '\v': case '\f':
|
---|
79 | case ' ': case '!': case '"': case '#': case '%':
|
---|
80 | case '&': case '\'': case '(': case ')': case '*':
|
---|
81 | case '+': case ',': case '-': case '.': case '/':
|
---|
82 | case '0': case '1': case '2': case '3': case '4':
|
---|
83 | case '5': case '6': case '7': case '8': case '9':
|
---|
84 | case ':': case ';': case '<': case '=': case '>':
|
---|
85 | case '?':
|
---|
86 | case 'A': case 'B': case 'C': case 'D': case 'E':
|
---|
87 | case 'F': case 'G': case 'H': case 'I': case 'J':
|
---|
88 | case 'K': case 'L': case 'M': case 'N': case 'O':
|
---|
89 | case 'P': case 'Q': case 'R': case 'S': case 'T':
|
---|
90 | case 'U': case 'V': case 'W': case 'X': case 'Y':
|
---|
91 | case 'Z':
|
---|
92 | case '[': case '\\': case ']': case '^': case '_':
|
---|
93 | case 'a': case 'b': case 'c': case 'd': case 'e':
|
---|
94 | case 'f': case 'g': case 'h': case 'i': case 'j':
|
---|
95 | case 'k': case 'l': case 'm': case 'n': case 'o':
|
---|
96 | case 'p': case 'q': case 'r': case 's': case 't':
|
---|
97 | case 'u': case 'v': case 'w': case 'x': case 'y':
|
---|
98 | case 'z': case '{': case '|': case '}': case '~':
|
---|
99 | /* c is in the ISO C "basic character set". */
|
---|
100 | buf[0] = c;
|
---|
101 | wc = (wchar_t) 0xBADFACE;
|
---|
102 | ret = mbrtowc (&wc, buf, 1, &state);
|
---|
103 | ASSERT (ret == 1);
|
---|
104 | ASSERT (wc == c);
|
---|
105 | ASSERT (mbsinit (&state));
|
---|
106 | ret = mbrtowc (NULL, buf, 1, &state);
|
---|
107 | ASSERT (ret == 1);
|
---|
108 | ASSERT (mbsinit (&state));
|
---|
109 | break;
|
---|
110 | }
|
---|
111 | }
|
---|
112 |
|
---|
113 | /* Test special calling convention, passing a NULL pointer. */
|
---|
114 | {
|
---|
115 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
116 | wc = (wchar_t) 0xBADFACE;
|
---|
117 | ret = mbrtowc (&wc, NULL, 5, &state);
|
---|
118 | ASSERT (ret == 0);
|
---|
119 | ASSERT (wc == (wchar_t) 0xBADFACE);
|
---|
120 | ASSERT (mbsinit (&state));
|
---|
121 | }
|
---|
122 |
|
---|
123 | if (argc > 1)
|
---|
124 | switch (argv[1][0])
|
---|
125 | {
|
---|
126 | case '1':
|
---|
127 | /* Locale encoding is ISO-8859-1 or ISO-8859-15. */
|
---|
128 | {
|
---|
129 | char input[] = "B\374\337er"; /* "Büßer" */
|
---|
130 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
131 |
|
---|
132 | wc = (wchar_t) 0xBADFACE;
|
---|
133 | ret = mbrtowc (&wc, input, 1, &state);
|
---|
134 | ASSERT (ret == 1);
|
---|
135 | ASSERT (wc == 'B');
|
---|
136 | ASSERT (mbsinit (&state));
|
---|
137 | input[0] = '\0';
|
---|
138 |
|
---|
139 | wc = (wchar_t) 0xBADFACE;
|
---|
140 | ret = mbrtowc (&wc, input + 1, 1, &state);
|
---|
141 | ASSERT (ret == 1);
|
---|
142 | ASSERT (wctob (wc) == (unsigned char) '\374');
|
---|
143 | ASSERT (mbsinit (&state));
|
---|
144 | input[1] = '\0';
|
---|
145 |
|
---|
146 | /* Test support of NULL first argument. */
|
---|
147 | ret = mbrtowc (NULL, input + 2, 3, &state);
|
---|
148 | ASSERT (ret == 1);
|
---|
149 | ASSERT (mbsinit (&state));
|
---|
150 |
|
---|
151 | wc = (wchar_t) 0xBADFACE;
|
---|
152 | ret = mbrtowc (&wc, input + 2, 3, &state);
|
---|
153 | ASSERT (ret == 1);
|
---|
154 | ASSERT (wctob (wc) == (unsigned char) '\337');
|
---|
155 | ASSERT (mbsinit (&state));
|
---|
156 | input[2] = '\0';
|
---|
157 |
|
---|
158 | wc = (wchar_t) 0xBADFACE;
|
---|
159 | ret = mbrtowc (&wc, input + 3, 2, &state);
|
---|
160 | ASSERT (ret == 1);
|
---|
161 | ASSERT (wc == 'e');
|
---|
162 | ASSERT (mbsinit (&state));
|
---|
163 | input[3] = '\0';
|
---|
164 |
|
---|
165 | wc = (wchar_t) 0xBADFACE;
|
---|
166 | ret = mbrtowc (&wc, input + 4, 1, &state);
|
---|
167 | ASSERT (ret == 1);
|
---|
168 | ASSERT (wc == 'r');
|
---|
169 | ASSERT (mbsinit (&state));
|
---|
170 | }
|
---|
171 | return 0;
|
---|
172 |
|
---|
173 | case '2':
|
---|
174 | /* Locale encoding is UTF-8. */
|
---|
175 | {
|
---|
176 | char input[] = "B\303\274\303\237er"; /* "Büßer" */
|
---|
177 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
178 |
|
---|
179 | wc = (wchar_t) 0xBADFACE;
|
---|
180 | ret = mbrtowc (&wc, input, 1, &state);
|
---|
181 | ASSERT (ret == 1);
|
---|
182 | ASSERT (wc == 'B');
|
---|
183 | ASSERT (mbsinit (&state));
|
---|
184 | input[0] = '\0';
|
---|
185 |
|
---|
186 | wc = (wchar_t) 0xBADFACE;
|
---|
187 | ret = mbrtowc (&wc, input + 1, 1, &state);
|
---|
188 | ASSERT (ret == (size_t)(-2));
|
---|
189 | ASSERT (wc == (wchar_t) 0xBADFACE);
|
---|
190 | ASSERT (!mbsinit (&state));
|
---|
191 | input[1] = '\0';
|
---|
192 |
|
---|
193 | wc = (wchar_t) 0xBADFACE;
|
---|
194 | ret = mbrtowc (&wc, input + 2, 5, &state);
|
---|
195 | ASSERT (ret == 1);
|
---|
196 | ASSERT (wctob (wc) == EOF);
|
---|
197 | ASSERT (mbsinit (&state));
|
---|
198 | input[2] = '\0';
|
---|
199 |
|
---|
200 | /* Test support of NULL first argument. */
|
---|
201 | ret = mbrtowc (NULL, input + 3, 4, &state);
|
---|
202 | ASSERT (ret == 2);
|
---|
203 | ASSERT (mbsinit (&state));
|
---|
204 |
|
---|
205 | wc = (wchar_t) 0xBADFACE;
|
---|
206 | ret = mbrtowc (&wc, input + 3, 4, &state);
|
---|
207 | ASSERT (ret == 2);
|
---|
208 | ASSERT (wctob (wc) == EOF);
|
---|
209 | ASSERT (mbsinit (&state));
|
---|
210 | input[3] = '\0';
|
---|
211 | input[4] = '\0';
|
---|
212 |
|
---|
213 | wc = (wchar_t) 0xBADFACE;
|
---|
214 | ret = mbrtowc (&wc, input + 5, 2, &state);
|
---|
215 | ASSERT (ret == 1);
|
---|
216 | ASSERT (wc == 'e');
|
---|
217 | ASSERT (mbsinit (&state));
|
---|
218 | input[5] = '\0';
|
---|
219 |
|
---|
220 | wc = (wchar_t) 0xBADFACE;
|
---|
221 | ret = mbrtowc (&wc, input + 6, 1, &state);
|
---|
222 | ASSERT (ret == 1);
|
---|
223 | ASSERT (wc == 'r');
|
---|
224 | ASSERT (mbsinit (&state));
|
---|
225 | }
|
---|
226 | return 0;
|
---|
227 |
|
---|
228 | case '3':
|
---|
229 | /* Locale encoding is EUC-JP. */
|
---|
230 | {
|
---|
231 | char input[] = "<\306\374\313\334\270\354>"; /* "<日本語>" */
|
---|
232 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
233 |
|
---|
234 | wc = (wchar_t) 0xBADFACE;
|
---|
235 | ret = mbrtowc (&wc, input, 1, &state);
|
---|
236 | ASSERT (ret == 1);
|
---|
237 | ASSERT (wc == '<');
|
---|
238 | ASSERT (mbsinit (&state));
|
---|
239 | input[0] = '\0';
|
---|
240 |
|
---|
241 | wc = (wchar_t) 0xBADFACE;
|
---|
242 | ret = mbrtowc (&wc, input + 1, 2, &state);
|
---|
243 | ASSERT (ret == 2);
|
---|
244 | ASSERT (wctob (wc) == EOF);
|
---|
245 | ASSERT (mbsinit (&state));
|
---|
246 | input[1] = '\0';
|
---|
247 | input[2] = '\0';
|
---|
248 |
|
---|
249 | wc = (wchar_t) 0xBADFACE;
|
---|
250 | ret = mbrtowc (&wc, input + 3, 1, &state);
|
---|
251 | ASSERT (ret == (size_t)(-2));
|
---|
252 | ASSERT (wc == (wchar_t) 0xBADFACE);
|
---|
253 | ASSERT (!mbsinit (&state));
|
---|
254 | input[3] = '\0';
|
---|
255 |
|
---|
256 | wc = (wchar_t) 0xBADFACE;
|
---|
257 | ret = mbrtowc (&wc, input + 4, 4, &state);
|
---|
258 | ASSERT (ret == 1);
|
---|
259 | ASSERT (wctob (wc) == EOF);
|
---|
260 | ASSERT (mbsinit (&state));
|
---|
261 | input[4] = '\0';
|
---|
262 |
|
---|
263 | /* Test support of NULL first argument. */
|
---|
264 | ret = mbrtowc (NULL, input + 5, 3, &state);
|
---|
265 | ASSERT (ret == 2);
|
---|
266 | ASSERT (mbsinit (&state));
|
---|
267 |
|
---|
268 | wc = (wchar_t) 0xBADFACE;
|
---|
269 | ret = mbrtowc (&wc, input + 5, 3, &state);
|
---|
270 | ASSERT (ret == 2);
|
---|
271 | ASSERT (wctob (wc) == EOF);
|
---|
272 | ASSERT (mbsinit (&state));
|
---|
273 | input[5] = '\0';
|
---|
274 | input[6] = '\0';
|
---|
275 |
|
---|
276 | wc = (wchar_t) 0xBADFACE;
|
---|
277 | ret = mbrtowc (&wc, input + 7, 1, &state);
|
---|
278 | ASSERT (ret == 1);
|
---|
279 | ASSERT (wc == '>');
|
---|
280 | ASSERT (mbsinit (&state));
|
---|
281 | }
|
---|
282 | return 0;
|
---|
283 |
|
---|
284 | case '4':
|
---|
285 | /* Locale encoding is GB18030. */
|
---|
286 | {
|
---|
287 | char input[] = "B\250\271\201\060\211\070er"; /* "Büßer" */
|
---|
288 | memset (&state, '\0', sizeof (mbstate_t));
|
---|
289 |
|
---|
290 | wc = (wchar_t) 0xBADFACE;
|
---|
291 | ret = mbrtowc (&wc, input, 1, &state);
|
---|
292 | ASSERT (ret == 1);
|
---|
293 | ASSERT (wc == 'B');
|
---|
294 | ASSERT (mbsinit (&state));
|
---|
295 | input[0] = '\0';
|
---|
296 |
|
---|
297 | wc = (wchar_t) 0xBADFACE;
|
---|
298 | ret = mbrtowc (&wc, input + 1, 1, &state);
|
---|
299 | ASSERT (ret == (size_t)(-2));
|
---|
300 | ASSERT (wc == (wchar_t) 0xBADFACE);
|
---|
301 | ASSERT (!mbsinit (&state));
|
---|
302 | input[1] = '\0';
|
---|
303 |
|
---|
304 | wc = (wchar_t) 0xBADFACE;
|
---|
305 | ret = mbrtowc (&wc, input + 2, 7, &state);
|
---|
306 | ASSERT (ret == 1);
|
---|
307 | ASSERT (wctob (wc) == EOF);
|
---|
308 | ASSERT (mbsinit (&state));
|
---|
309 | input[2] = '\0';
|
---|
310 |
|
---|
311 | /* Test support of NULL first argument. */
|
---|
312 | ret = mbrtowc (NULL, input + 3, 6, &state);
|
---|
313 | ASSERT (ret == 4);
|
---|
314 | ASSERT (mbsinit (&state));
|
---|
315 |
|
---|
316 | wc = (wchar_t) 0xBADFACE;
|
---|
317 | ret = mbrtowc (&wc, input + 3, 6, &state);
|
---|
318 | ASSERT (ret == 4);
|
---|
319 | ASSERT (wctob (wc) == EOF);
|
---|
320 | ASSERT (mbsinit (&state));
|
---|
321 | input[3] = '\0';
|
---|
322 | input[4] = '\0';
|
---|
323 | input[5] = '\0';
|
---|
324 | input[6] = '\0';
|
---|
325 |
|
---|
326 | wc = (wchar_t) 0xBADFACE;
|
---|
327 | ret = mbrtowc (&wc, input + 7, 2, &state);
|
---|
328 | ASSERT (ret == 1);
|
---|
329 | ASSERT (wc == 'e');
|
---|
330 | ASSERT (mbsinit (&state));
|
---|
331 | input[5] = '\0';
|
---|
332 |
|
---|
333 | wc = (wchar_t) 0xBADFACE;
|
---|
334 | ret = mbrtowc (&wc, input + 8, 1, &state);
|
---|
335 | ASSERT (ret == 1);
|
---|
336 | ASSERT (wc == 'r');
|
---|
337 | ASSERT (mbsinit (&state));
|
---|
338 | }
|
---|
339 | return 0;
|
---|
340 | }
|
---|
341 |
|
---|
342 | return 1;
|
---|
343 | }
|
---|