1 /* bug 19727: Testing UTF conversions with UTF16 surrogates as input.
2 Copyright (C) 2016-2022 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4
5 The GNU C Library is free software; you can redistribute it and/or
6 modify it under the terms of the GNU Lesser General Public
7 License as published by the Free Software Foundation; either
8 version 2.1 of the License, or (at your option) any later version.
9
10 The GNU C Library is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 Lesser General Public License for more details.
14
15 You should have received a copy of the GNU Lesser General Public
16 License along with the GNU C Library; if not, see
17 <https://www.gnu.org/licenses/>. */
18
19 #include <stdio.h>
20 #include <stdlib.h>
21 #include <errno.h>
22 #include <string.h>
23 #include <inttypes.h>
24 #include <iconv.h>
25 #include <byteswap.h>
26
27 static int
run_conversion(const char * from,const char * to,char * inbuf,size_t inbuflen,int exp_errno,int line)28 run_conversion (const char *from, const char *to, char *inbuf, size_t inbuflen,
29 int exp_errno, int line)
30 {
31 char outbuf[16];
32 iconv_t cd;
33 char *inptr;
34 size_t inlen;
35 char *outptr;
36 size_t outlen;
37 size_t n;
38 int e;
39 int fails = 0;
40
41 cd = iconv_open (to, from);
42 if (cd == (iconv_t) -1)
43 {
44 printf ("line %d: cannot convert from %s to %s: %m\n", line, from, to);
45 return 1;
46 }
47
48 inptr = (char *) inbuf;
49 inlen = inbuflen;
50 outptr = outbuf;
51 outlen = sizeof (outbuf);
52
53 errno = 0;
54 n = iconv (cd, &inptr, &inlen, &outptr, &outlen);
55 e = errno;
56
57 if (exp_errno == 0)
58 {
59 if (n == (size_t) -1)
60 {
61 puts ("n should be >= 0, but n == -1");
62 fails ++;
63 }
64
65 if (e != 0)
66 {
67 printf ("errno should be 0: 'Success', but errno == %d: '%s'\n"
68 , e, strerror(e));
69 fails ++;
70 }
71 }
72 else
73 {
74 if (n != (size_t) -1)
75 {
76 printf ("n should be -1, but n == %zd\n", n);
77 fails ++;
78 }
79
80 if (e != exp_errno)
81 {
82 printf ("errno should be %d: '%s', but errno == %d: '%s'\n"
83 , exp_errno, strerror (exp_errno), e, strerror (e));
84 fails ++;
85 }
86 }
87
88 iconv_close (cd);
89
90 if (fails > 0)
91 {
92 printf ("Errors in line %d while converting %s to %s.\n\n"
93 , line, from, to);
94 }
95
96 return fails;
97 }
98
99 static int
do_test(void)100 do_test (void)
101 {
102 int fails = 0;
103 char buf[4];
104
105 /* This test runs iconv() with UTF character in range of an UTF16 surrogate.
106 UTF-16 high surrogate is in range 0xD800..0xDBFF and
107 UTF-16 low surrogate is in range 0xDC00..0xDFFF.
108 Converting from or to UTF-xx has to report errors in those cases.
109 In UTF-16, surrogate pairs with a high surrogate in front of a low
110 surrogate is valid. */
111
112 /* Use RUN_UCS4_UTF32_INPUT to test conversion ...
113
114 ... from INTERNAL to UTF-xx[LE|BE]:
115 Converting from UCS4 to UTF-xx[LE|BE] first converts UCS4 to INTERNAL
116 without checking for UTF-16 surrogate values
117 and then converts from INTERNAL to UTF-xx[LE|BE].
118 The latter conversion has to report an error in those cases.
119
120 ... from UTF-32[LE|BE] to INTERNAL:
121 Converting directly from UTF-32LE to UTF-8|16 is needed,
122 because e.g. s390x has iconv-modules which converts directly. */
123 #define RUN_UCS4_UTF32_INPUT(b0, b1, b2, b3, err, line) \
124 buf[0] = b0; \
125 buf[1] = b1; \
126 buf[2] = b2; \
127 buf[3] = b3; \
128 fails += run_conversion ("UCS4", "UTF-8", buf, 4, err, line); \
129 fails += run_conversion ("UCS4", "UTF-16LE", buf, 4, err, line); \
130 fails += run_conversion ("UCS4", "UTF-16BE", buf, 4, err, line); \
131 fails += run_conversion ("UCS4", "UTF-32LE", buf, 4, err, line); \
132 fails += run_conversion ("UCS4", "UTF-32BE", buf, 4, err, line); \
133 fails += run_conversion ("UTF-32BE", "WCHAR_T", buf, 4, err, line); \
134 fails += run_conversion ("UTF-32BE", "UTF-8", buf, 4, err, line); \
135 fails += run_conversion ("UTF-32BE", "UTF-16LE", buf, 4, err, line); \
136 fails += run_conversion ("UTF-32BE", "UTF-16BE", buf, 4, err, line); \
137 buf[0] = b3; \
138 buf[1] = b2; \
139 buf[2] = b1; \
140 buf[3] = b0; \
141 fails += run_conversion ("UTF-32LE", "WCHAR_T", buf, 4, err, line); \
142 fails += run_conversion ("UTF-32LE", "UTF-8", buf, 4, err, line); \
143 fails += run_conversion ("UTF-32LE", "UTF-16LE", buf, 4, err, line); \
144 fails += run_conversion ("UTF-32LE", "UTF-16BE", buf, 4, err, line);
145
146 /* Use UCS4/UTF32 input of 0xD7FF. */
147 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xD7, 0xFF, 0, __LINE__);
148
149 /* Use UCS4/UTF32 input of 0xD800. */
150 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xD8, 0x00, EILSEQ, __LINE__);
151
152 /* Use UCS4/UTF32 input of 0xDBFF. */
153 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xDB, 0xFF, EILSEQ, __LINE__);
154
155 /* Use UCS4/UTF32 input of 0xDC00. */
156 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xDC, 0x00, EILSEQ, __LINE__);
157
158 /* Use UCS4/UTF32 input of 0xDFFF. */
159 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xDF, 0xFF, EILSEQ, __LINE__);
160
161 /* Use UCS4/UTF32 input of 0xE000. */
162 RUN_UCS4_UTF32_INPUT (0x0, 0x0, 0xE0, 0x00, 0, __LINE__);
163
164
165 /* Use RUN_UTF16_INPUT to test conversion from UTF16[LE|BE] to INTERNAL.
166 Converting directly from UTF-16 to UTF-8|32 is needed,
167 because e.g. s390x has iconv-modules which converts directly.
168 Use len == 2 or 4 to specify one or two UTF-16 characters. */
169 #define RUN_UTF16_INPUT(b0, b1, b2, b3, len, err, line) \
170 buf[0] = b0; \
171 buf[1] = b1; \
172 buf[2] = b2; \
173 buf[3] = b3; \
174 fails += run_conversion ("UTF-16BE", "WCHAR_T", buf, len, err, line); \
175 fails += run_conversion ("UTF-16BE", "UTF-8", buf, len, err, line); \
176 fails += run_conversion ("UTF-16BE", "UTF-32LE", buf, len, err, line); \
177 fails += run_conversion ("UTF-16BE", "UTF-32BE", buf, len, err, line); \
178 buf[0] = b1; \
179 buf[1] = b0; \
180 buf[2] = b3; \
181 buf[3] = b2; \
182 fails += run_conversion ("UTF-16LE", "WCHAR_T", buf, len, err, line); \
183 fails += run_conversion ("UTF-16LE", "UTF-8", buf, len, err, line); \
184 fails += run_conversion ("UTF-16LE", "UTF-32LE", buf, len, err, line); \
185 fails += run_conversion ("UTF-16LE", "UTF-32BE", buf, len, err, line);
186
187 /* Use UTF16 input of 0xD7FF. */
188 RUN_UTF16_INPUT (0xD7, 0xFF, 0xD7, 0xFF, 4, 0, __LINE__);
189
190 /* Use [single] UTF16 high surrogate 0xD800 [with a valid character behind].
191 And check an UTF16 surrogate pair [without valid low surrogate]. */
192 RUN_UTF16_INPUT (0xD8, 0x0, 0x0, 0x0, 2, EINVAL, __LINE__);
193 RUN_UTF16_INPUT (0xD8, 0x0, 0xD7, 0xFF, 4, EILSEQ, __LINE__);
194 RUN_UTF16_INPUT (0xD8, 0x0, 0xD8, 0x0, 4, EILSEQ, __LINE__);
195 RUN_UTF16_INPUT (0xD8, 0x0, 0xE0, 0x0, 4, EILSEQ, __LINE__);
196 RUN_UTF16_INPUT (0xD8, 0x0, 0xDC, 0x0, 4, 0, __LINE__);
197
198 /* Use [single] UTF16 high surrogate 0xDBFF [with a valid character behind].
199 And check an UTF16 surrogate pair [without valid low surrogate]. */
200 RUN_UTF16_INPUT (0xDB, 0xFF, 0x0, 0x0, 2, EINVAL, __LINE__);
201 RUN_UTF16_INPUT (0xDB, 0xFF, 0xD7, 0xFF, 4, EILSEQ, __LINE__);
202 RUN_UTF16_INPUT (0xDB, 0xFF, 0xDB, 0xFF, 4, EILSEQ, __LINE__);
203 RUN_UTF16_INPUT (0xDB, 0xFF, 0xE0, 0x0, 4, EILSEQ, __LINE__);
204 RUN_UTF16_INPUT (0xDB, 0xFF, 0xDF, 0xFF, 4, 0, __LINE__);
205
206 /* Use single UTF16 low surrogate 0xDC00 [with a valid character behind].
207 And check an UTF16 surrogate pair [without valid high surrogate]. */
208 RUN_UTF16_INPUT (0xDC, 0x0, 0x0, 0x0, 2, EILSEQ, __LINE__);
209 RUN_UTF16_INPUT (0xDC, 0x0, 0xD7, 0xFF, 4, EILSEQ, __LINE__);
210 RUN_UTF16_INPUT (0xD8, 0x0, 0xDC, 0x0, 4, 0, __LINE__);
211 RUN_UTF16_INPUT (0xD7, 0xFF, 0xDC, 0x0, 4, EILSEQ, __LINE__);
212 RUN_UTF16_INPUT (0xDC, 0x0, 0xDC, 0x0, 4, EILSEQ, __LINE__);
213 RUN_UTF16_INPUT (0xE0, 0x0, 0xDC, 0x0, 4, EILSEQ, __LINE__);
214
215 /* Use single UTF16 low surrogate 0xDFFF [with a valid character behind].
216 And check an UTF16 surrogate pair [without valid high surrogate]. */
217 RUN_UTF16_INPUT (0xDF, 0xFF, 0x0, 0x0, 2, EILSEQ, __LINE__);
218 RUN_UTF16_INPUT (0xDF, 0xFF, 0xD7, 0xFF, 4, EILSEQ, __LINE__);
219 RUN_UTF16_INPUT (0xDB, 0xFF, 0xDF, 0xFF, 4, 0, __LINE__);
220 RUN_UTF16_INPUT (0xD7, 0xFF, 0xDF, 0xFF, 4, EILSEQ, __LINE__);
221 RUN_UTF16_INPUT (0xDF, 0xFF, 0xDF, 0xFF, 4, EILSEQ, __LINE__);
222 RUN_UTF16_INPUT (0xE0, 0x0, 0xDF, 0xFF, 4, EILSEQ, __LINE__);
223
224 /* Use UCS4/UTF32 input of 0xE000. */
225 RUN_UTF16_INPUT (0xE0, 0x0, 0xE0, 0x0, 4, 0, __LINE__);
226
227
228 /* Use RUN_UTF8_3BYTE_INPUT to test conversion from UTF-8 to INTERNAL.
229 Converting directly from UTF-8 to UTF-16|32 is needed,
230 because e.g. s390x has iconv-modules which converts directly. */
231 #define RUN_UTF8_3BYTE_INPUT(b0, b1, b2, err, line) \
232 buf[0] = b0; \
233 buf[1] = b1; \
234 buf[2] = b2; \
235 fails += run_conversion ("UTF-8", "WCHAR_T", buf, 3, err, line); \
236 fails += run_conversion ("UTF-8", "UTF-16LE", buf, 3, err, line); \
237 fails += run_conversion ("UTF-8", "UTF-16BE", buf, 3, err, line); \
238 fails += run_conversion ("UTF-8", "UTF-32LE", buf, 3, err, line); \
239 fails += run_conversion ("UTF-8", "UTF-32BE", buf, 3, err, line);
240
241 /* Use UTF-8 input of 0xD7FF. */
242 RUN_UTF8_3BYTE_INPUT (0xED, 0x9F, 0xBF, 0, __LINE__);
243
244 /* Use UTF-8 input of 0xD800. */
245 RUN_UTF8_3BYTE_INPUT (0xED, 0xA0, 0x80, EILSEQ, __LINE__);
246
247 /* Use UTF-8 input of 0xDBFF. */
248 RUN_UTF8_3BYTE_INPUT (0xED, 0xAF, 0xBF, EILSEQ, __LINE__);
249
250 /* Use UTF-8 input of 0xDC00. */
251 RUN_UTF8_3BYTE_INPUT (0xED, 0xB0, 0x80, EILSEQ, __LINE__);
252
253 /* Use UTF-8 input of 0xDFFF. */
254 RUN_UTF8_3BYTE_INPUT (0xED, 0xBF, 0xBF, EILSEQ, __LINE__);
255
256 /* Use UTF-8 input of 0xF000. */
257 RUN_UTF8_3BYTE_INPUT (0xEF, 0x80, 0x80, 0, __LINE__);
258
259 return fails > 0 ? EXIT_FAILURE : EXIT_SUCCESS;
260 }
261
262 #define TEST_FUNCTION do_test ()
263 #include "../test-skeleton.c"
264