]>
Commit | Line | Data |
---|---|---|
422c46a8 AL |
1 | /* |
2 | * Copyright IBM, Corp. 2009 | |
e549e716 | 3 | * Copyright (c) 2013, 2015 Red Hat Inc. |
422c46a8 AL |
4 | * |
5 | * Authors: | |
6 | * Anthony Liguori <aliguori@us.ibm.com> | |
d6244e2c | 7 | * Markus Armbruster <armbru@redhat.com> |
422c46a8 AL |
8 | * |
9 | * This work is licensed under the terms of the GNU LGPL, version 2.1 or later. | |
10 | * See the COPYING.LIB file in the top-level directory. | |
11 | * | |
12 | */ | |
aec4b054 | 13 | |
681c28a3 | 14 | #include "qemu/osdep.h" |
422c46a8 | 15 | |
aec4b054 | 16 | #include "qapi/error.h" |
6b673957 | 17 | #include "qapi/qmp/qbool.h" |
7b1b5d19 | 18 | #include "qapi/qmp/qjson.h" |
28035bcd | 19 | #include "qapi/qmp/qlit.h" |
15280c36 MA |
20 | #include "qapi/qmp/qnull.h" |
21 | #include "qapi/qmp/qnum.h" | |
fc81fa1e | 22 | #include "qapi/qmp/qstring.h" |
5f454e66 | 23 | #include "qemu/unicode.h" |
422c46a8 AL |
24 | #include "qemu-common.h" |
25 | ||
4e1df9b7 MA |
26 | static QString *from_json_str(const char *jstr, bool single, Error **errp) |
27 | { | |
28 | char quote = single ? '\'' : '"'; | |
29 | char *qjstr = g_strdup_printf("%c%s%c", quote, jstr, quote); | |
30 | QString *ret = qobject_to(QString, qobject_from_json(qjstr, errp)); | |
31 | ||
32 | g_free(qjstr); | |
33 | return ret; | |
34 | } | |
35 | ||
36 | static char *to_json_str(QString *str) | |
37 | { | |
38 | QString *json = qobject_to_json(QOBJECT(str)); | |
39 | char *jstr; | |
40 | ||
41 | if (!json) { | |
42 | return NULL; | |
43 | } | |
44 | /* peel off double quotes */ | |
45 | jstr = g_strndup(qstring_get_str(json) + 1, | |
46 | qstring_get_length(json) - 2); | |
47 | qobject_unref(json); | |
48 | return jstr; | |
49 | } | |
50 | ||
ef76dc59 | 51 | static void escaped_string(void) |
422c46a8 | 52 | { |
422c46a8 | 53 | struct { |
4e1df9b7 MA |
54 | /* Content of JSON string to parse with qobject_from_json() */ |
55 | const char *json_in; | |
56 | /* Expected parse output; to unparse with qobject_to_json() */ | |
57 | const char *utf8_out; | |
6ee59202 | 58 | int skip; |
422c46a8 | 59 | } test_cases[] = { |
f3cfdd3a | 60 | { "\\b\\f\\n\\r\\t\\\\\\\"", "\b\f\n\r\t\\\"" }, |
e0fe2a97 | 61 | { "\\/\\'", "/'", .skip = 1 }, |
4e1df9b7 MA |
62 | { "single byte utf-8 \\u0020", "single byte utf-8 ", .skip = 1 }, |
63 | { "double byte utf-8 \\u00A2", "double byte utf-8 \xc2\xa2" }, | |
64 | { "triple byte utf-8 \\u20AC", "triple byte utf-8 \xe2\x82\xac" }, | |
e0fe2a97 MA |
65 | { "quadruple byte utf-8 \\uD834\\uDD1E", /* U+1D11E */ |
66 | /* bug: want \xF0\x9D\x84\x9E */ | |
67 | "quadruple byte utf-8 \xED\xA0\xB4\xED\xB4\x9E", .skip = 1 }, | |
68 | { "\\", NULL }, | |
69 | { "\\z", NULL }, | |
70 | { "\\ux", NULL }, | |
71 | { "\\u1x", NULL }, | |
72 | { "\\u12x", NULL }, | |
73 | { "\\u123x", NULL }, | |
74 | { "\\u12345", "\341\210\2645" }, | |
75 | { "\\u0000x", "x", .skip = 1}, /* bug: want \xC0\x80x */ | |
76 | { "unpaired leading surrogate \\uD800", | |
77 | /* bug: not rejected */ | |
78 | "unpaired leading surrogate \355\240\200", .skip = 1 }, | |
79 | { "unpaired leading surrogate \\uD800\\uCAFE", | |
80 | /* bug: not rejected */ | |
81 | "unpaired leading surrogate \355\240\200\354\253\276", .skip = 1 }, | |
82 | { "unpaired leading surrogate \\uD800\\uD801\\uDC02", | |
83 | /* bug: not rejected */ | |
84 | "unpaired leading surrogate \355\240\200\355\240\201\355\260\202", | |
85 | .skip = 1 }, | |
86 | { "unpaired trailing surrogate \\uDC00", | |
87 | /* bug: not rejected */ | |
88 | "unpaired trailing surrogate \355\260\200", .skip = 1}, | |
89 | { "backward surrogate pair \\uDC00\\uD800", | |
90 | /* bug: not rejected */ | |
91 | "backward surrogate pair \355\260\200\355\240\200", .skip = 1}, | |
92 | { "noncharacter U+FDD0 \\uFDD0", | |
93 | /* bug: not rejected */ | |
94 | "noncharacter U+FDD0 \xEF\xB7\x90", .skip = 1}, | |
95 | { "noncharacter U+FDEF \\uFDEF", | |
96 | /* bug: not rejected */ | |
97 | "noncharacter U+FDEF \xEF\xB7\xAF", .skip = 1}, | |
98 | { "noncharacter U+1FFFE \\uD87F\\uDFFE", | |
99 | /* bug: not rejected */ | |
100 | "noncharacter U+1FFFE \xED\xA1\xBF\xED\xBF\xBE", .skip = 1}, | |
101 | { "noncharacter U+10FFFF \\uDC3F\\uDFFF", | |
102 | /* bug: not rejected */ | |
103 | "noncharacter U+10FFFF \xED\xB0\xBF\xED\xBF\xBF", .skip = 1}, | |
422c46a8 AL |
104 | {} |
105 | }; | |
4e1df9b7 MA |
106 | int i, j; |
107 | QString *cstr; | |
108 | char *jstr; | |
422c46a8 | 109 | |
4e1df9b7 MA |
110 | for (i = 0; test_cases[i].json_in; i++) { |
111 | for (j = 0; j < 2; j++) { | |
e0fe2a97 MA |
112 | if (test_cases[i].utf8_out) { |
113 | cstr = from_json_str(test_cases[i].json_in, j, &error_abort); | |
114 | g_assert_cmpstr(qstring_get_try_str(cstr), | |
115 | ==, test_cases[i].utf8_out); | |
116 | if (!test_cases[i].skip) { | |
117 | jstr = to_json_str(cstr); | |
118 | g_assert_cmpstr(jstr, ==, test_cases[i].json_in); | |
119 | g_free(jstr); | |
120 | } | |
121 | qobject_unref(cstr); | |
122 | } else { | |
123 | cstr = from_json_str(test_cases[i].json_in, j, NULL); | |
124 | g_assert(!cstr); | |
4e1df9b7 | 125 | } |
6ee59202 | 126 | } |
422c46a8 AL |
127 | } |
128 | } | |
422c46a8 | 129 | |
069946f4 | 130 | static void string_with_quotes(void) |
422c46a8 | 131 | { |
069946f4 MA |
132 | const char *test_cases[] = { |
133 | "\"the bee's knees\"", | |
134 | "'double quote \"'", | |
135 | NULL | |
422c46a8 | 136 | }; |
422c46a8 | 137 | int i; |
069946f4 MA |
138 | QString *str; |
139 | char *cstr; | |
422c46a8 | 140 | |
069946f4 MA |
141 | for (i = 0; test_cases[i]; i++) { |
142 | str = qobject_to(QString, | |
143 | qobject_from_json(test_cases[i], &error_abort)); | |
363e13f8 | 144 | g_assert(str); |
069946f4 MA |
145 | cstr = g_strndup(test_cases[i] + 1, strlen(test_cases[i]) - 2); |
146 | g_assert_cmpstr(qstring_get_str(str), ==, cstr); | |
147 | g_free(cstr); | |
cb3e7f08 | 148 | qobject_unref(str); |
422c46a8 AL |
149 | } |
150 | } | |
422c46a8 | 151 | |
3960c41f MA |
152 | static void utf8_string(void) |
153 | { | |
154 | /* | |
155 | * FIXME Current behavior for invalid UTF-8 sequences is | |
156 | * incorrect. This test expects current, incorrect results. | |
157 | * They're all marked "bug:" below, and are to be replaced by | |
158 | * correct ones as the bugs get fixed. | |
159 | * | |
160 | * The JSON parser rejects some invalid sequences, but accepts | |
161 | * others without correcting the problem. | |
162 | * | |
e2ec3f97 MA |
163 | * We should either reject all invalid sequences, or minimize |
164 | * overlong sequences and replace all other invalid sequences by a | |
165 | * suitable replacement character. A common choice for | |
166 | * replacement is U+FFFD. | |
3960c41f MA |
167 | * |
168 | * Problem: we can't easily deal with embedded U+0000. Parsing | |
169 | * the JSON string "this \\u0000" is fun" yields "this \0 is fun", | |
170 | * which gets misinterpreted as NUL-terminated "this ". We should | |
171 | * consider using overlong encoding \xC0\x80 for U+0000 ("modified | |
172 | * UTF-8"). | |
173 | * | |
1d50c8e9 | 174 | * Most test cases are scraped from Markus Kuhn's UTF-8 decoder |
3960c41f MA |
175 | * capability and stress test at |
176 | * http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt | |
177 | */ | |
178 | static const struct { | |
6ad8444f | 179 | /* Content of JSON string to parse with qobject_from_json() */ |
3960c41f | 180 | const char *json_in; |
6ad8444f | 181 | /* Expected parse output */ |
3960c41f | 182 | const char *utf8_out; |
6ad8444f MA |
183 | /* Expected unparse output, defaults to @json_in */ |
184 | const char *json_out; | |
3960c41f MA |
185 | } test_cases[] = { |
186 | /* | |
187 | * Bug markers used here: | |
188 | * - bug: not corrected | |
189 | * JSON parser fails to correct invalid sequence(s) | |
190 | * - bug: rejected | |
191 | * JSON parser rejects invalid sequence(s) | |
192 | * We may choose to define this as feature | |
3960c41f MA |
193 | */ |
194 | ||
6bc93a34 MA |
195 | /* 0 Control characters */ |
196 | { | |
197 | /* | |
198 | * Note: \x00 is impossible, other representations of | |
199 | * U+0000 are covered under 4.3 | |
200 | */ | |
201 | "\x01\x02\x03\x04\x05\x06\x07" | |
202 | "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" | |
203 | "\x10\x11\x12\x13\x14\x15\x16\x17" | |
204 | "\x18\x19\x1A\x1B\x1C\x1D\x1E\x1F", | |
205 | /* bug: not corrected (valid UTF-8, but invalid JSON) */ | |
206 | "\x01\x02\x03\x04\x05\x06\x07" | |
207 | "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" | |
208 | "\x10\x11\x12\x13\x14\x15\x16\x17" | |
209 | "\x18\x19\x1A\x1B\x1C\x1D\x1E\x1F", | |
210 | "\\u0001\\u0002\\u0003\\u0004\\u0005\\u0006\\u0007" | |
211 | "\\b\\t\\n\\u000B\\f\\r\\u000E\\u000F" | |
212 | "\\u0010\\u0011\\u0012\\u0013\\u0014\\u0015\\u0016\\u0017" | |
213 | "\\u0018\\u0019\\u001A\\u001B\\u001C\\u001D\\u001E\\u001F", | |
214 | }, | |
3960c41f MA |
215 | /* 1 Some correct UTF-8 text */ |
216 | { | |
217 | /* a bit of German */ | |
3960c41f MA |
218 | "Falsches \xC3\x9C" "ben von Xylophonmusik qu\xC3\xA4lt" |
219 | " jeden gr\xC3\xB6\xC3\x9F" "eren Zwerg.", | |
6ad8444f MA |
220 | "Falsches \xC3\x9C" "ben von Xylophonmusik qu\xC3\xA4lt" |
221 | " jeden gr\xC3\xB6\xC3\x9F" "eren Zwerg.", | |
222 | "Falsches \\u00DCben von Xylophonmusik qu\\u00E4lt" | |
223 | " jeden gr\\u00F6\\u00DFeren Zwerg.", | |
3960c41f MA |
224 | }, |
225 | { | |
226 | /* a bit of Greek */ | |
3960c41f | 227 | "\xCE\xBA\xE1\xBD\xB9\xCF\x83\xCE\xBC\xCE\xB5", |
6ad8444f MA |
228 | "\xCE\xBA\xE1\xBD\xB9\xCF\x83\xCE\xBC\xCE\xB5", |
229 | "\\u03BA\\u1F79\\u03C3\\u03BC\\u03B5", | |
3960c41f MA |
230 | }, |
231 | /* 2 Boundary condition test cases */ | |
232 | /* 2.1 First possible sequence of a certain length */ | |
32846e93 | 233 | /* |
6bc93a34 MA |
234 | * 2.1.1 1 byte U+0020 |
235 | * Control characters are already covered by their own test | |
236 | * case under 0. Test the first 1 byte non-control character | |
237 | * here. | |
32846e93 | 238 | */ |
3960c41f | 239 | { |
6bc93a34 MA |
240 | " ", |
241 | " ", | |
3960c41f MA |
242 | }, |
243 | /* 2.1.2 2 bytes U+0080 */ | |
244 | { | |
3960c41f | 245 | "\xC2\x80", |
6ad8444f MA |
246 | "\xC2\x80", |
247 | "\\u0080", | |
3960c41f MA |
248 | }, |
249 | /* 2.1.3 3 bytes U+0800 */ | |
250 | { | |
3960c41f | 251 | "\xE0\xA0\x80", |
6ad8444f MA |
252 | "\xE0\xA0\x80", |
253 | "\\u0800", | |
3960c41f MA |
254 | }, |
255 | /* 2.1.4 4 bytes U+10000 */ | |
256 | { | |
3960c41f | 257 | "\xF0\x90\x80\x80", |
6ad8444f MA |
258 | "\xF0\x90\x80\x80", |
259 | "\\uD800\\uDC00", | |
3960c41f MA |
260 | }, |
261 | /* 2.1.5 5 bytes U+200000 */ | |
262 | { | |
6ad8444f | 263 | "\xF8\x88\x80\x80\x80", |
e2ec3f97 | 264 | NULL, /* bug: rejected */ |
6ad8444f | 265 | "\\uFFFD", |
3960c41f MA |
266 | }, |
267 | /* 2.1.6 6 bytes U+4000000 */ | |
268 | { | |
6ad8444f | 269 | "\xFC\x84\x80\x80\x80\x80", |
e2ec3f97 | 270 | NULL, /* bug: rejected */ |
6ad8444f | 271 | "\\uFFFD", |
3960c41f MA |
272 | }, |
273 | /* 2.2 Last possible sequence of a certain length */ | |
274 | /* 2.2.1 1 byte U+007F */ | |
275 | { | |
3960c41f | 276 | "\x7F", |
6ad8444f MA |
277 | "\x7F", |
278 | "\\u007F", | |
3960c41f MA |
279 | }, |
280 | /* 2.2.2 2 bytes U+07FF */ | |
281 | { | |
3960c41f | 282 | "\xDF\xBF", |
6ad8444f MA |
283 | "\xDF\xBF", |
284 | "\\u07FF", | |
3960c41f | 285 | }, |
1d50c8e9 MA |
286 | /* |
287 | * 2.2.3 3 bytes U+FFFC | |
288 | * The last possible sequence is actually U+FFFF. But that's | |
289 | * a noncharacter, and already covered by its own test case | |
290 | * under 5.3. Same for U+FFFE. U+FFFD is the last character | |
291 | * in the BMP, and covered under 2.3. Because of U+FFFD's | |
292 | * special role as replacement character, it's worth testing | |
293 | * U+FFFC here. | |
294 | */ | |
3960c41f | 295 | { |
1d50c8e9 | 296 | "\xEF\xBF\xBC", |
6ad8444f MA |
297 | "\xEF\xBF\xBC", |
298 | "\\uFFFC", | |
3960c41f MA |
299 | }, |
300 | /* 2.2.4 4 bytes U+1FFFFF */ | |
301 | { | |
6ad8444f | 302 | "\xF7\xBF\xBF\xBF", |
e2ec3f97 | 303 | NULL, /* bug: rejected */ |
6ad8444f | 304 | "\\uFFFD", |
3960c41f MA |
305 | }, |
306 | /* 2.2.5 5 bytes U+3FFFFFF */ | |
307 | { | |
6ad8444f | 308 | "\xFB\xBF\xBF\xBF\xBF", |
e2ec3f97 | 309 | NULL, /* bug: rejected */ |
6ad8444f | 310 | "\\uFFFD", |
3960c41f MA |
311 | }, |
312 | /* 2.2.6 6 bytes U+7FFFFFFF */ | |
313 | { | |
6ad8444f | 314 | "\xFD\xBF\xBF\xBF\xBF\xBF", |
e2ec3f97 | 315 | NULL, /* bug: rejected */ |
6ad8444f | 316 | "\\uFFFD", |
3960c41f MA |
317 | }, |
318 | /* 2.3 Other boundary conditions */ | |
319 | { | |
d6244e2c | 320 | /* last one before surrogate range: U+D7FF */ |
3960c41f | 321 | "\xED\x9F\xBF", |
6ad8444f MA |
322 | "\xED\x9F\xBF", |
323 | "\\uD7FF", | |
3960c41f MA |
324 | }, |
325 | { | |
d6244e2c | 326 | /* first one after surrogate range: U+E000 */ |
3960c41f | 327 | "\xEE\x80\x80", |
6ad8444f MA |
328 | "\xEE\x80\x80", |
329 | "\\uE000", | |
3960c41f MA |
330 | }, |
331 | { | |
d6244e2c | 332 | /* last one in BMP: U+FFFD */ |
3960c41f | 333 | "\xEF\xBF\xBD", |
6ad8444f MA |
334 | "\xEF\xBF\xBD", |
335 | "\\uFFFD", | |
3960c41f MA |
336 | }, |
337 | { | |
1d50c8e9 | 338 | /* last one in last plane: U+10FFFD */ |
1d50c8e9 | 339 | "\xF4\x8F\xBF\xBD", |
6ad8444f MA |
340 | "\xF4\x8F\xBF\xBD", |
341 | "\\uDBFF\\uDFFD" | |
3960c41f MA |
342 | }, |
343 | { | |
d6244e2c | 344 | /* first one beyond Unicode range: U+110000 */ |
3960c41f | 345 | "\xF4\x90\x80\x80", |
6ad8444f MA |
346 | "\xF4\x90\x80\x80", |
347 | "\\uFFFD", | |
3960c41f MA |
348 | }, |
349 | /* 3 Malformed sequences */ | |
350 | /* 3.1 Unexpected continuation bytes */ | |
351 | /* 3.1.1 First continuation byte */ | |
352 | { | |
6ad8444f | 353 | "\x80", |
3960c41f | 354 | "\x80", /* bug: not corrected */ |
6ad8444f | 355 | "\\uFFFD", |
3960c41f MA |
356 | }, |
357 | /* 3.1.2 Last continuation byte */ | |
358 | { | |
6ad8444f | 359 | "\xBF", |
3960c41f | 360 | "\xBF", /* bug: not corrected */ |
6ad8444f | 361 | "\\uFFFD", |
3960c41f MA |
362 | }, |
363 | /* 3.1.3 2 continuation bytes */ | |
364 | { | |
6ad8444f | 365 | "\x80\xBF", |
3960c41f | 366 | "\x80\xBF", /* bug: not corrected */ |
6ad8444f | 367 | "\\uFFFD\\uFFFD", |
3960c41f MA |
368 | }, |
369 | /* 3.1.4 3 continuation bytes */ | |
370 | { | |
6ad8444f | 371 | "\x80\xBF\x80", |
3960c41f | 372 | "\x80\xBF\x80", /* bug: not corrected */ |
6ad8444f | 373 | "\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
374 | }, |
375 | /* 3.1.5 4 continuation bytes */ | |
376 | { | |
6ad8444f | 377 | "\x80\xBF\x80\xBF", |
3960c41f | 378 | "\x80\xBF\x80\xBF", /* bug: not corrected */ |
6ad8444f | 379 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
380 | }, |
381 | /* 3.1.6 5 continuation bytes */ | |
382 | { | |
6ad8444f | 383 | "\x80\xBF\x80\xBF\x80", |
3960c41f | 384 | "\x80\xBF\x80\xBF\x80", /* bug: not corrected */ |
6ad8444f | 385 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
386 | }, |
387 | /* 3.1.7 6 continuation bytes */ | |
388 | { | |
6ad8444f | 389 | "\x80\xBF\x80\xBF\x80\xBF", |
3960c41f | 390 | "\x80\xBF\x80\xBF\x80\xBF", /* bug: not corrected */ |
6ad8444f | 391 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
392 | }, |
393 | /* 3.1.8 7 continuation bytes */ | |
394 | { | |
6ad8444f | 395 | "\x80\xBF\x80\xBF\x80\xBF\x80", |
3960c41f | 396 | "\x80\xBF\x80\xBF\x80\xBF\x80", /* bug: not corrected */ |
6ad8444f | 397 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
398 | }, |
399 | /* 3.1.9 Sequence of all 64 possible continuation bytes */ | |
400 | { | |
6ad8444f | 401 | "\x80\x81\x82\x83\x84\x85\x86\x87" |
3960c41f MA |
402 | "\x88\x89\x8A\x8B\x8C\x8D\x8E\x8F" |
403 | "\x90\x91\x92\x93\x94\x95\x96\x97" | |
404 | "\x98\x99\x9A\x9B\x9C\x9D\x9E\x9F" | |
405 | "\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7" | |
406 | "\xA8\xA9\xAA\xAB\xAC\xAD\xAE\xAF" | |
407 | "\xB0\xB1\xB2\xB3\xB4\xB5\xB6\xB7" | |
6ad8444f | 408 | "\xB8\xB9\xBA\xBB\xBC\xBD\xBE\xBF", |
3960c41f MA |
409 | /* bug: not corrected */ |
410 | "\x80\x81\x82\x83\x84\x85\x86\x87" | |
411 | "\x88\x89\x8A\x8B\x8C\x8D\x8E\x8F" | |
412 | "\x90\x91\x92\x93\x94\x95\x96\x97" | |
413 | "\x98\x99\x9A\x9B\x9C\x9D\x9E\x9F" | |
414 | "\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7" | |
415 | "\xA8\xA9\xAA\xAB\xAC\xAD\xAE\xAF" | |
416 | "\xB0\xB1\xB2\xB3\xB4\xB5\xB6\xB7" | |
417 | "\xB8\xB9\xBA\xBB\xBC\xBD\xBE\xBF", | |
e2ec3f97 MA |
418 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" |
419 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
420 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
421 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
422 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
423 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
6ad8444f MA |
424 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" |
425 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
3960c41f MA |
426 | }, |
427 | /* 3.2 Lonely start characters */ | |
428 | /* 3.2.1 All 32 first bytes of 2-byte sequences, followed by space */ | |
429 | { | |
6ad8444f | 430 | "\xC0 \xC1 \xC2 \xC3 \xC4 \xC5 \xC6 \xC7 " |
3960c41f MA |
431 | "\xC8 \xC9 \xCA \xCB \xCC \xCD \xCE \xCF " |
432 | "\xD0 \xD1 \xD2 \xD3 \xD4 \xD5 \xD6 \xD7 " | |
6ad8444f | 433 | "\xD8 \xD9 \xDA \xDB \xDC \xDD \xDE \xDF ", |
5f454e66 | 434 | NULL, /* bug: rejected (partly, see FIXME below) */ |
e2ec3f97 MA |
435 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD " |
436 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD " | |
6ad8444f MA |
437 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD " |
438 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD ", | |
3960c41f MA |
439 | }, |
440 | /* 3.2.2 All 16 first bytes of 3-byte sequences, followed by space */ | |
441 | { | |
6ad8444f MA |
442 | "\xE0 \xE1 \xE2 \xE3 \xE4 \xE5 \xE6 \xE7 " |
443 | "\xE8 \xE9 \xEA \xEB \xEC \xED \xEE \xEF ", | |
3960c41f MA |
444 | /* bug: not corrected */ |
445 | "\xE0 \xE1 \xE2 \xE3 \xE4 \xE5 \xE6 \xE7 " | |
446 | "\xE8 \xE9 \xEA \xEB \xEC \xED \xEE \xEF ", | |
6ad8444f MA |
447 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD " |
448 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD ", | |
3960c41f MA |
449 | }, |
450 | /* 3.2.3 All 8 first bytes of 4-byte sequences, followed by space */ | |
451 | { | |
6ad8444f | 452 | "\xF0 \xF1 \xF2 \xF3 \xF4 \xF5 \xF6 \xF7 ", |
5f454e66 | 453 | NULL, /* bug: rejected (partly, see FIXME below) */ |
6ad8444f | 454 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD \\uFFFD ", |
3960c41f MA |
455 | }, |
456 | /* 3.2.4 All 4 first bytes of 5-byte sequences, followed by space */ | |
457 | { | |
6ad8444f | 458 | "\xF8 \xF9 \xFA \xFB ", |
3960c41f | 459 | NULL, /* bug: rejected */ |
6ad8444f | 460 | "\\uFFFD \\uFFFD \\uFFFD \\uFFFD ", |
3960c41f MA |
461 | }, |
462 | /* 3.2.5 All 2 first bytes of 6-byte sequences, followed by space */ | |
463 | { | |
6ad8444f | 464 | "\xFC \xFD ", |
3960c41f | 465 | NULL, /* bug: rejected */ |
6ad8444f | 466 | "\\uFFFD \\uFFFD ", |
3960c41f MA |
467 | }, |
468 | /* 3.3 Sequences with last continuation byte missing */ | |
469 | /* 3.3.1 2-byte sequence with last byte missing (U+0000) */ | |
470 | { | |
6ad8444f | 471 | "\xC0", |
3960c41f | 472 | NULL, /* bug: rejected */ |
6ad8444f | 473 | "\\uFFFD", |
3960c41f MA |
474 | }, |
475 | /* 3.3.2 3-byte sequence with last byte missing (U+0000) */ | |
476 | { | |
6ad8444f | 477 | "\xE0\x80", |
3960c41f | 478 | "\xE0\x80", /* bug: not corrected */ |
6ad8444f | 479 | "\\uFFFD", |
3960c41f MA |
480 | }, |
481 | /* 3.3.3 4-byte sequence with last byte missing (U+0000) */ | |
482 | { | |
6ad8444f | 483 | "\xF0\x80\x80", |
3960c41f | 484 | "\xF0\x80\x80", /* bug: not corrected */ |
6ad8444f | 485 | "\\uFFFD", |
3960c41f MA |
486 | }, |
487 | /* 3.3.4 5-byte sequence with last byte missing (U+0000) */ | |
488 | { | |
6ad8444f | 489 | "\xF8\x80\x80\x80", |
3960c41f | 490 | NULL, /* bug: rejected */ |
6ad8444f | 491 | "\\uFFFD", |
3960c41f MA |
492 | }, |
493 | /* 3.3.5 6-byte sequence with last byte missing (U+0000) */ | |
494 | { | |
6ad8444f | 495 | "\xFC\x80\x80\x80\x80", |
3960c41f | 496 | NULL, /* bug: rejected */ |
6ad8444f | 497 | "\\uFFFD", |
3960c41f MA |
498 | }, |
499 | /* 3.3.6 2-byte sequence with last byte missing (U+07FF) */ | |
500 | { | |
6ad8444f | 501 | "\xDF", |
3960c41f | 502 | "\xDF", /* bug: not corrected */ |
6ad8444f | 503 | "\\uFFFD", |
3960c41f MA |
504 | }, |
505 | /* 3.3.7 3-byte sequence with last byte missing (U+FFFF) */ | |
506 | { | |
6ad8444f | 507 | "\xEF\xBF", |
3960c41f | 508 | "\xEF\xBF", /* bug: not corrected */ |
6ad8444f | 509 | "\\uFFFD", |
3960c41f MA |
510 | }, |
511 | /* 3.3.8 4-byte sequence with last byte missing (U+1FFFFF) */ | |
512 | { | |
6ad8444f | 513 | "\xF7\xBF\xBF", |
3960c41f | 514 | NULL, /* bug: rejected */ |
6ad8444f | 515 | "\\uFFFD", |
3960c41f MA |
516 | }, |
517 | /* 3.3.9 5-byte sequence with last byte missing (U+3FFFFFF) */ | |
518 | { | |
6ad8444f | 519 | "\xFB\xBF\xBF\xBF", |
3960c41f | 520 | NULL, /* bug: rejected */ |
6ad8444f | 521 | "\\uFFFD", |
3960c41f MA |
522 | }, |
523 | /* 3.3.10 6-byte sequence with last byte missing (U+7FFFFFFF) */ | |
524 | { | |
6ad8444f | 525 | "\xFD\xBF\xBF\xBF\xBF", |
3960c41f | 526 | NULL, /* bug: rejected */ |
6ad8444f | 527 | "\\uFFFD", |
3960c41f MA |
528 | }, |
529 | /* 3.4 Concatenation of incomplete sequences */ | |
530 | { | |
6ad8444f MA |
531 | "\xC0\xE0\x80\xF0\x80\x80\xF8\x80\x80\x80\xFC\x80\x80\x80\x80" |
532 | "\xDF\xEF\xBF\xF7\xBF\xBF\xFB\xBF\xBF\xBF\xFD\xBF\xBF\xBF\xBF", | |
5f454e66 | 533 | NULL, /* bug: rejected (partly, see FIXME below) */ |
6ad8444f MA |
534 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" |
535 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD", | |
3960c41f MA |
536 | }, |
537 | /* 3.5 Impossible bytes */ | |
538 | { | |
6ad8444f | 539 | "\xFE", |
3960c41f | 540 | NULL, /* bug: rejected */ |
6ad8444f | 541 | "\\uFFFD", |
3960c41f MA |
542 | }, |
543 | { | |
6ad8444f | 544 | "\xFF", |
3960c41f | 545 | NULL, /* bug: rejected */ |
6ad8444f | 546 | "\\uFFFD", |
3960c41f MA |
547 | }, |
548 | { | |
6ad8444f | 549 | "\xFE\xFE\xFF\xFF", |
3960c41f | 550 | NULL, /* bug: rejected */ |
6ad8444f | 551 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD", |
3960c41f MA |
552 | }, |
553 | /* 4 Overlong sequences */ | |
554 | /* 4.1 Overlong '/' */ | |
555 | { | |
6ad8444f | 556 | "\xC0\xAF", |
3960c41f | 557 | NULL, /* bug: rejected */ |
6ad8444f | 558 | "\\uFFFD", |
3960c41f MA |
559 | }, |
560 | { | |
6ad8444f | 561 | "\xE0\x80\xAF", |
3960c41f | 562 | "\xE0\x80\xAF", /* bug: not corrected */ |
6ad8444f | 563 | "\\uFFFD", |
3960c41f MA |
564 | }, |
565 | { | |
6ad8444f | 566 | "\xF0\x80\x80\xAF", |
3960c41f | 567 | "\xF0\x80\x80\xAF", /* bug: not corrected */ |
6ad8444f | 568 | "\\uFFFD", |
3960c41f MA |
569 | }, |
570 | { | |
6ad8444f | 571 | "\xF8\x80\x80\x80\xAF", |
3960c41f | 572 | NULL, /* bug: rejected */ |
6ad8444f | 573 | "\\uFFFD", |
3960c41f MA |
574 | }, |
575 | { | |
6ad8444f | 576 | "\xFC\x80\x80\x80\x80\xAF", |
3960c41f | 577 | NULL, /* bug: rejected */ |
6ad8444f | 578 | "\\uFFFD", |
3960c41f | 579 | }, |
d6244e2c MA |
580 | /* |
581 | * 4.2 Maximum overlong sequences | |
582 | * Highest Unicode value that is still resulting in an | |
583 | * overlong sequence if represented with the given number of | |
584 | * bytes. This is a boundary test for safe UTF-8 decoders. | |
585 | */ | |
3960c41f MA |
586 | { |
587 | /* \U+007F */ | |
6ad8444f | 588 | "\xC1\xBF", |
3960c41f | 589 | NULL, /* bug: rejected */ |
6ad8444f | 590 | "\\uFFFD", |
3960c41f MA |
591 | }, |
592 | { | |
593 | /* \U+07FF */ | |
6ad8444f | 594 | "\xE0\x9F\xBF", |
3960c41f | 595 | "\xE0\x9F\xBF", /* bug: not corrected */ |
6ad8444f | 596 | "\\uFFFD", |
3960c41f MA |
597 | }, |
598 | { | |
1d50c8e9 MA |
599 | /* |
600 | * \U+FFFC | |
601 | * The actual maximum would be U+FFFF, but that's a | |
602 | * noncharacter. Testing U+FFFC seems more useful. See | |
603 | * also 2.2.3 | |
604 | */ | |
6ad8444f | 605 | "\xF0\x8F\xBF\xBC", |
1d50c8e9 | 606 | "\xF0\x8F\xBF\xBC", /* bug: not corrected */ |
6ad8444f | 607 | "\\uFFFD", |
3960c41f MA |
608 | }, |
609 | { | |
610 | /* \U+1FFFFF */ | |
6ad8444f | 611 | "\xF8\x87\xBF\xBF\xBF", |
3960c41f | 612 | NULL, /* bug: rejected */ |
6ad8444f | 613 | "\\uFFFD", |
3960c41f MA |
614 | }, |
615 | { | |
616 | /* \U+3FFFFFF */ | |
6ad8444f | 617 | "\xFC\x83\xBF\xBF\xBF\xBF", |
3960c41f | 618 | NULL, /* bug: rejected */ |
6ad8444f | 619 | "\\uFFFD", |
3960c41f MA |
620 | }, |
621 | /* 4.3 Overlong representation of the NUL character */ | |
622 | { | |
623 | /* \U+0000 */ | |
6ad8444f | 624 | "\xC0\x80", |
3960c41f | 625 | NULL, /* bug: rejected */ |
6ad8444f | 626 | "\\u0000", |
3960c41f MA |
627 | }, |
628 | { | |
629 | /* \U+0000 */ | |
6ad8444f | 630 | "\xE0\x80\x80", |
3960c41f | 631 | "\xE0\x80\x80", /* bug: not corrected */ |
6ad8444f | 632 | "\\uFFFD", |
3960c41f MA |
633 | }, |
634 | { | |
635 | /* \U+0000 */ | |
6ad8444f | 636 | "\xF0\x80\x80\x80", |
3960c41f | 637 | "\xF0\x80\x80\x80", /* bug: not corrected */ |
6ad8444f | 638 | "\\uFFFD", |
3960c41f MA |
639 | }, |
640 | { | |
641 | /* \U+0000 */ | |
6ad8444f | 642 | "\xF8\x80\x80\x80\x80", |
3960c41f | 643 | NULL, /* bug: rejected */ |
6ad8444f | 644 | "\\uFFFD", |
3960c41f MA |
645 | }, |
646 | { | |
647 | /* \U+0000 */ | |
6ad8444f | 648 | "\xFC\x80\x80\x80\x80\x80", |
3960c41f | 649 | NULL, /* bug: rejected */ |
6ad8444f | 650 | "\\uFFFD", |
3960c41f MA |
651 | }, |
652 | /* 5 Illegal code positions */ | |
653 | /* 5.1 Single UTF-16 surrogates */ | |
654 | { | |
655 | /* \U+D800 */ | |
6ad8444f | 656 | "\xED\xA0\x80", |
3960c41f | 657 | "\xED\xA0\x80", /* bug: not corrected */ |
6ad8444f | 658 | "\\uFFFD", |
3960c41f MA |
659 | }, |
660 | { | |
661 | /* \U+DB7F */ | |
6ad8444f | 662 | "\xED\xAD\xBF", |
3960c41f | 663 | "\xED\xAD\xBF", /* bug: not corrected */ |
6ad8444f | 664 | "\\uFFFD", |
3960c41f MA |
665 | }, |
666 | { | |
667 | /* \U+DB80 */ | |
6ad8444f | 668 | "\xED\xAE\x80", |
3960c41f | 669 | "\xED\xAE\x80", /* bug: not corrected */ |
6ad8444f | 670 | "\\uFFFD", |
3960c41f MA |
671 | }, |
672 | { | |
673 | /* \U+DBFF */ | |
6ad8444f | 674 | "\xED\xAF\xBF", |
3960c41f | 675 | "\xED\xAF\xBF", /* bug: not corrected */ |
6ad8444f | 676 | "\\uFFFD", |
3960c41f MA |
677 | }, |
678 | { | |
679 | /* \U+DC00 */ | |
6ad8444f | 680 | "\xED\xB0\x80", |
3960c41f | 681 | "\xED\xB0\x80", /* bug: not corrected */ |
6ad8444f | 682 | "\\uFFFD", |
3960c41f MA |
683 | }, |
684 | { | |
685 | /* \U+DF80 */ | |
6ad8444f | 686 | "\xED\xBE\x80", |
3960c41f | 687 | "\xED\xBE\x80", /* bug: not corrected */ |
6ad8444f | 688 | "\\uFFFD", |
3960c41f MA |
689 | }, |
690 | { | |
691 | /* \U+DFFF */ | |
6ad8444f | 692 | "\xED\xBF\xBF", |
3960c41f | 693 | "\xED\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 694 | "\\uFFFD", |
3960c41f MA |
695 | }, |
696 | /* 5.2 Paired UTF-16 surrogates */ | |
697 | { | |
698 | /* \U+D800\U+DC00 */ | |
6ad8444f | 699 | "\xED\xA0\x80\xED\xB0\x80", |
3960c41f | 700 | "\xED\xA0\x80\xED\xB0\x80", /* bug: not corrected */ |
6ad8444f | 701 | "\\uFFFD\\uFFFD", |
3960c41f MA |
702 | }, |
703 | { | |
704 | /* \U+D800\U+DFFF */ | |
6ad8444f | 705 | "\xED\xA0\x80\xED\xBF\xBF", |
3960c41f | 706 | "\xED\xA0\x80\xED\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 707 | "\\uFFFD\\uFFFD", |
3960c41f MA |
708 | }, |
709 | { | |
710 | /* \U+DB7F\U+DC00 */ | |
6ad8444f | 711 | "\xED\xAD\xBF\xED\xB0\x80", |
3960c41f | 712 | "\xED\xAD\xBF\xED\xB0\x80", /* bug: not corrected */ |
6ad8444f | 713 | "\\uFFFD\\uFFFD", |
3960c41f MA |
714 | }, |
715 | { | |
716 | /* \U+DB7F\U+DFFF */ | |
6ad8444f | 717 | "\xED\xAD\xBF\xED\xBF\xBF", |
3960c41f | 718 | "\xED\xAD\xBF\xED\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 719 | "\\uFFFD\\uFFFD", |
3960c41f MA |
720 | }, |
721 | { | |
722 | /* \U+DB80\U+DC00 */ | |
6ad8444f | 723 | "\xED\xAE\x80\xED\xB0\x80", |
3960c41f | 724 | "\xED\xAE\x80\xED\xB0\x80", /* bug: not corrected */ |
6ad8444f | 725 | "\\uFFFD\\uFFFD", |
3960c41f MA |
726 | }, |
727 | { | |
728 | /* \U+DB80\U+DFFF */ | |
6ad8444f | 729 | "\xED\xAE\x80\xED\xBF\xBF", |
3960c41f | 730 | "\xED\xAE\x80\xED\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 731 | "\\uFFFD\\uFFFD", |
3960c41f MA |
732 | }, |
733 | { | |
734 | /* \U+DBFF\U+DC00 */ | |
6ad8444f | 735 | "\xED\xAF\xBF\xED\xB0\x80", |
3960c41f | 736 | "\xED\xAF\xBF\xED\xB0\x80", /* bug: not corrected */ |
6ad8444f | 737 | "\\uFFFD\\uFFFD", |
3960c41f MA |
738 | }, |
739 | { | |
740 | /* \U+DBFF\U+DFFF */ | |
6ad8444f | 741 | "\xED\xAF\xBF\xED\xBF\xBF", |
3960c41f | 742 | "\xED\xAF\xBF\xED\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 743 | "\\uFFFD\\uFFFD", |
3960c41f MA |
744 | }, |
745 | /* 5.3 Other illegal code positions */ | |
1d50c8e9 | 746 | /* BMP noncharacters */ |
3960c41f MA |
747 | { |
748 | /* \U+FFFE */ | |
6ad8444f | 749 | "\xEF\xBF\xBE", |
3960c41f | 750 | "\xEF\xBF\xBE", /* bug: not corrected */ |
6ad8444f | 751 | "\\uFFFD", |
3960c41f MA |
752 | }, |
753 | { | |
754 | /* \U+FFFF */ | |
6ad8444f | 755 | "\xEF\xBF\xBF", |
3960c41f | 756 | "\xEF\xBF\xBF", /* bug: not corrected */ |
6ad8444f | 757 | "\\uFFFD", |
3960c41f | 758 | }, |
1d50c8e9 MA |
759 | { |
760 | /* U+FDD0 */ | |
6ad8444f | 761 | "\xEF\xB7\x90", |
1d50c8e9 | 762 | "\xEF\xB7\x90", /* bug: not corrected */ |
6ad8444f | 763 | "\\uFFFD", |
1d50c8e9 MA |
764 | }, |
765 | { | |
766 | /* U+FDEF */ | |
6ad8444f | 767 | "\xEF\xB7\xAF", |
1d50c8e9 | 768 | "\xEF\xB7\xAF", /* bug: not corrected */ |
6ad8444f | 769 | "\\uFFFD", |
1d50c8e9 MA |
770 | }, |
771 | /* Plane 1 .. 16 noncharacters */ | |
772 | { | |
773 | /* U+1FFFE U+1FFFF U+2FFFE U+2FFFF ... U+10FFFE U+10FFFF */ | |
6ad8444f | 774 | "\xF0\x9F\xBF\xBE\xF0\x9F\xBF\xBF" |
1d50c8e9 MA |
775 | "\xF0\xAF\xBF\xBE\xF0\xAF\xBF\xBF" |
776 | "\xF0\xBF\xBF\xBE\xF0\xBF\xBF\xBF" | |
777 | "\xF1\x8F\xBF\xBE\xF1\x8F\xBF\xBF" | |
778 | "\xF1\x9F\xBF\xBE\xF1\x9F\xBF\xBF" | |
779 | "\xF1\xAF\xBF\xBE\xF1\xAF\xBF\xBF" | |
780 | "\xF1\xBF\xBF\xBE\xF1\xBF\xBF\xBF" | |
781 | "\xF2\x8F\xBF\xBE\xF2\x8F\xBF\xBF" | |
782 | "\xF2\x9F\xBF\xBE\xF2\x9F\xBF\xBF" | |
783 | "\xF2\xAF\xBF\xBE\xF2\xAF\xBF\xBF" | |
784 | "\xF2\xBF\xBF\xBE\xF2\xBF\xBF\xBF" | |
785 | "\xF3\x8F\xBF\xBE\xF3\x8F\xBF\xBF" | |
786 | "\xF3\x9F\xBF\xBE\xF3\x9F\xBF\xBF" | |
787 | "\xF3\xAF\xBF\xBE\xF3\xAF\xBF\xBF" | |
788 | "\xF3\xBF\xBF\xBE\xF3\xBF\xBF\xBF" | |
6ad8444f | 789 | "\xF4\x8F\xBF\xBE\xF4\x8F\xBF\xBF", |
1d50c8e9 MA |
790 | /* bug: not corrected */ |
791 | "\xF0\x9F\xBF\xBE\xF0\x9F\xBF\xBF" | |
792 | "\xF0\xAF\xBF\xBE\xF0\xAF\xBF\xBF" | |
793 | "\xF0\xBF\xBF\xBE\xF0\xBF\xBF\xBF" | |
794 | "\xF1\x8F\xBF\xBE\xF1\x8F\xBF\xBF" | |
795 | "\xF1\x9F\xBF\xBE\xF1\x9F\xBF\xBF" | |
796 | "\xF1\xAF\xBF\xBE\xF1\xAF\xBF\xBF" | |
797 | "\xF1\xBF\xBF\xBE\xF1\xBF\xBF\xBF" | |
798 | "\xF2\x8F\xBF\xBE\xF2\x8F\xBF\xBF" | |
799 | "\xF2\x9F\xBF\xBE\xF2\x9F\xBF\xBF" | |
800 | "\xF2\xAF\xBF\xBE\xF2\xAF\xBF\xBF" | |
801 | "\xF2\xBF\xBF\xBE\xF2\xBF\xBF\xBF" | |
802 | "\xF3\x8F\xBF\xBE\xF3\x8F\xBF\xBF" | |
803 | "\xF3\x9F\xBF\xBE\xF3\x9F\xBF\xBF" | |
804 | "\xF3\xAF\xBF\xBE\xF3\xAF\xBF\xBF" | |
805 | "\xF3\xBF\xBF\xBE\xF3\xBF\xBF\xBF" | |
806 | "\xF4\x8F\xBF\xBE\xF4\x8F\xBF\xBF", | |
e2ec3f97 MA |
807 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" |
808 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" | |
6ad8444f MA |
809 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD" |
810 | "\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD\\uFFFD", | |
1d50c8e9 | 811 | }, |
3960c41f MA |
812 | {} |
813 | }; | |
6ad8444f | 814 | int i, j; |
3960c41f | 815 | QString *str; |
5f454e66 MA |
816 | const char *json_in, *utf8_out, *utf8_in, *json_out, *tail; |
817 | char *end, *in, *jstr; | |
3960c41f MA |
818 | |
819 | for (i = 0; test_cases[i].json_in; i++) { | |
6ad8444f MA |
820 | for (j = 0; j < 2; j++) { |
821 | json_in = test_cases[i].json_in; | |
822 | utf8_out = test_cases[i].utf8_out; | |
32846e93 | 823 | utf8_in = test_cases[i].utf8_out ?: test_cases[i].json_in; |
6ad8444f | 824 | json_out = test_cases[i].json_out ?: test_cases[i].json_in; |
3960c41f | 825 | |
6ad8444f MA |
826 | /* Parse @json_in, expect @utf8_out */ |
827 | if (utf8_out) { | |
828 | str = from_json_str(json_in, j, &error_abort); | |
829 | g_assert_cmpstr(qstring_get_try_str(str), ==, utf8_out); | |
830 | qobject_unref(str); | |
831 | } else { | |
832 | str = from_json_str(json_in, j, NULL); | |
833 | g_assert(!str); | |
5f454e66 MA |
834 | /* |
835 | * Failure may be due to any sequence, but *all* sequences | |
836 | * are expected to fail. Test each one in isolation. | |
837 | */ | |
838 | for (tail = json_in; *tail; tail = end) { | |
839 | mod_utf8_codepoint(tail, 6, &end); | |
840 | if (*end == ' ') { | |
841 | end++; | |
842 | } | |
843 | in = strndup(tail, end - tail); | |
844 | str = from_json_str(in, j, NULL); | |
845 | /* | |
846 | * FIXME JSON parser accepts invalid sequence | |
847 | * starting with \xC2..\xF4 | |
848 | */ | |
849 | if (*in >= '\xC2' && *in <= '\xF4') { | |
850 | g_free(str); | |
851 | str = NULL; | |
852 | } | |
853 | g_assert(!str); | |
854 | g_free(in); | |
855 | } | |
6ad8444f | 856 | } |
3960c41f | 857 | |
6ad8444f MA |
858 | /* Unparse @utf8_in, expect @json_out */ |
859 | str = qstring_from_str(utf8_in); | |
860 | jstr = to_json_str(str); | |
861 | g_assert_cmpstr(jstr, ==, json_out); | |
862 | qobject_unref(str); | |
863 | g_free(jstr); | |
3960c41f | 864 | |
6ad8444f MA |
865 | /* |
866 | * Parse @json_out right back | |
867 | * Disabled, because qobject_from_json() is buggy, and I can't | |
868 | * be bothered to add the expected incorrect results. | |
869 | * FIXME Enable once these bugs have been fixed. | |
870 | */ | |
871 | if (0 && json_out != json_in) { | |
872 | str = from_json_str(json_out, j, &error_abort); | |
873 | g_assert_cmpstr(qstring_get_try_str(str), ==, utf8_out); | |
874 | } | |
3960c41f MA |
875 | } |
876 | } | |
877 | } | |
878 | ||
ef76dc59 | 879 | static void simple_number(void) |
422c46a8 AL |
880 | { |
881 | int i; | |
882 | struct { | |
883 | const char *encoded; | |
884 | int64_t decoded; | |
6ee59202 | 885 | int skip; |
422c46a8 AL |
886 | } test_cases[] = { |
887 | { "0", 0 }, | |
888 | { "1234", 1234 }, | |
889 | { "1", 1 }, | |
890 | { "-32", -32 }, | |
6ee59202 | 891 | { "-0", 0, .skip = 1 }, |
422c46a8 AL |
892 | { }, |
893 | }; | |
894 | ||
895 | for (i = 0; test_cases[i].encoded; i++) { | |
01b2ffce MAL |
896 | QNum *qnum; |
897 | int64_t val; | |
422c46a8 | 898 | |
7dc847eb HR |
899 | qnum = qobject_to(QNum, |
900 | qobject_from_json(test_cases[i].encoded, | |
901 | &error_abort)); | |
01b2ffce MAL |
902 | g_assert(qnum); |
903 | g_assert(qnum_get_try_int(qnum, &val)); | |
904 | g_assert_cmpint(val, ==, test_cases[i].decoded); | |
6ee59202 AL |
905 | if (test_cases[i].skip == 0) { |
906 | QString *str; | |
907 | ||
01b2ffce | 908 | str = qobject_to_json(QOBJECT(qnum)); |
ef76dc59 | 909 | g_assert(strcmp(qstring_get_str(str), test_cases[i].encoded) == 0); |
cb3e7f08 | 910 | qobject_unref(str); |
6ee59202 | 911 | } |
422c46a8 | 912 | |
cb3e7f08 | 913 | qobject_unref(qnum); |
422c46a8 AL |
914 | } |
915 | } | |
422c46a8 | 916 | |
2bc7cfea MAL |
917 | static void large_number(void) |
918 | { | |
919 | const char *maxu64 = "18446744073709551615"; /* 2^64-1 */ | |
920 | const char *gtu64 = "18446744073709551616"; /* 2^64 */ | |
921 | const char *lti64 = "-9223372036854775809"; /* -2^63 - 1 */ | |
922 | QNum *qnum; | |
923 | QString *str; | |
924 | uint64_t val; | |
925 | int64_t ival; | |
926 | ||
7dc847eb | 927 | qnum = qobject_to(QNum, qobject_from_json(maxu64, &error_abort)); |
2bc7cfea MAL |
928 | g_assert(qnum); |
929 | g_assert_cmpuint(qnum_get_uint(qnum), ==, 18446744073709551615U); | |
930 | g_assert(!qnum_get_try_int(qnum, &ival)); | |
931 | ||
932 | str = qobject_to_json(QOBJECT(qnum)); | |
933 | g_assert_cmpstr(qstring_get_str(str), ==, maxu64); | |
cb3e7f08 MAL |
934 | qobject_unref(str); |
935 | qobject_unref(qnum); | |
2bc7cfea | 936 | |
7dc847eb | 937 | qnum = qobject_to(QNum, qobject_from_json(gtu64, &error_abort)); |
2bc7cfea MAL |
938 | g_assert(qnum); |
939 | g_assert_cmpfloat(qnum_get_double(qnum), ==, 18446744073709552e3); | |
940 | g_assert(!qnum_get_try_uint(qnum, &val)); | |
941 | g_assert(!qnum_get_try_int(qnum, &ival)); | |
942 | ||
943 | str = qobject_to_json(QOBJECT(qnum)); | |
944 | g_assert_cmpstr(qstring_get_str(str), ==, gtu64); | |
cb3e7f08 MAL |
945 | qobject_unref(str); |
946 | qobject_unref(qnum); | |
2bc7cfea | 947 | |
7dc847eb | 948 | qnum = qobject_to(QNum, qobject_from_json(lti64, &error_abort)); |
2bc7cfea MAL |
949 | g_assert(qnum); |
950 | g_assert_cmpfloat(qnum_get_double(qnum), ==, -92233720368547758e2); | |
951 | g_assert(!qnum_get_try_uint(qnum, &val)); | |
952 | g_assert(!qnum_get_try_int(qnum, &ival)); | |
953 | ||
954 | str = qobject_to_json(QOBJECT(qnum)); | |
955 | g_assert_cmpstr(qstring_get_str(str), ==, "-9223372036854775808"); | |
cb3e7f08 MAL |
956 | qobject_unref(str); |
957 | qobject_unref(qnum); | |
2bc7cfea MAL |
958 | } |
959 | ||
ef76dc59 | 960 | static void float_number(void) |
422c46a8 AL |
961 | { |
962 | int i; | |
963 | struct { | |
964 | const char *encoded; | |
965 | double decoded; | |
6ee59202 | 966 | int skip; |
422c46a8 AL |
967 | } test_cases[] = { |
968 | { "32.43", 32.43 }, | |
969 | { "0.222", 0.222 }, | |
970 | { "-32.12313", -32.12313 }, | |
6ee59202 | 971 | { "-32.20e-10", -32.20e-10, .skip = 1 }, |
422c46a8 AL |
972 | { }, |
973 | }; | |
974 | ||
975 | for (i = 0; test_cases[i].encoded; i++) { | |
976 | QObject *obj; | |
01b2ffce | 977 | QNum *qnum; |
422c46a8 | 978 | |
aec4b054 | 979 | obj = qobject_from_json(test_cases[i].encoded, &error_abort); |
7dc847eb | 980 | qnum = qobject_to(QNum, obj); |
01b2ffce MAL |
981 | g_assert(qnum); |
982 | g_assert(qnum_get_double(qnum) == test_cases[i].decoded); | |
422c46a8 | 983 | |
6ee59202 AL |
984 | if (test_cases[i].skip == 0) { |
985 | QString *str; | |
986 | ||
987 | str = qobject_to_json(obj); | |
ef76dc59 | 988 | g_assert(strcmp(qstring_get_str(str), test_cases[i].encoded) == 0); |
cb3e7f08 | 989 | qobject_unref(str); |
6ee59202 AL |
990 | } |
991 | ||
cb3e7f08 | 992 | qobject_unref(qnum); |
422c46a8 AL |
993 | } |
994 | } | |
422c46a8 | 995 | |
ef76dc59 | 996 | static void keyword_literal(void) |
422c46a8 AL |
997 | { |
998 | QObject *obj; | |
999 | QBool *qbool; | |
006ca09f | 1000 | QNull *null; |
6ee59202 | 1001 | QString *str; |
422c46a8 | 1002 | |
aec4b054 | 1003 | obj = qobject_from_json("true", &error_abort); |
7dc847eb | 1004 | qbool = qobject_to(QBool, obj); |
dfad9ec4 | 1005 | g_assert(qbool); |
fc48ffc3 | 1006 | g_assert(qbool_get_bool(qbool) == true); |
422c46a8 | 1007 | |
6ee59202 | 1008 | str = qobject_to_json(obj); |
ef76dc59 | 1009 | g_assert(strcmp(qstring_get_str(str), "true") == 0); |
cb3e7f08 | 1010 | qobject_unref(str); |
6ee59202 | 1011 | |
cb3e7f08 | 1012 | qobject_unref(qbool); |
422c46a8 | 1013 | |
aec4b054 | 1014 | obj = qobject_from_json("false", &error_abort); |
7dc847eb | 1015 | qbool = qobject_to(QBool, obj); |
dfad9ec4 | 1016 | g_assert(qbool); |
fc48ffc3 | 1017 | g_assert(qbool_get_bool(qbool) == false); |
422c46a8 | 1018 | |
6ee59202 | 1019 | str = qobject_to_json(obj); |
ef76dc59 | 1020 | g_assert(strcmp(qstring_get_str(str), "false") == 0); |
cb3e7f08 | 1021 | qobject_unref(str); |
6ee59202 | 1022 | |
cb3e7f08 | 1023 | qobject_unref(qbool); |
422c46a8 | 1024 | |
2e933f57 MA |
1025 | obj = qobject_from_json("null", &error_abort); |
1026 | g_assert(obj != NULL); | |
1027 | g_assert(qobject_type(obj) == QTYPE_QNULL); | |
1028 | ||
1029 | null = qnull(); | |
1030 | g_assert(QOBJECT(null) == obj); | |
1031 | ||
1032 | qobject_unref(obj); | |
1033 | qobject_unref(null); | |
1034 | } | |
1035 | ||
1036 | static void interpolation_valid(void) | |
1037 | { | |
1038 | long long value_lld = 0x123456789abcdefLL; | |
1039 | long value_ld = (long)value_lld; | |
1040 | int value_d = (int)value_lld; | |
1041 | unsigned long long value_llu = 0xfedcba9876543210ULL; | |
1042 | unsigned long value_lu = (unsigned long)value_llu; | |
1043 | unsigned value_u = (unsigned)value_llu; | |
1044 | double value_f = 2.323423423; | |
1045 | const char *value_s = "hello world"; | |
1046 | QObject *value_p = QOBJECT(qnull()); | |
1047 | QBool *qbool; | |
1048 | QNum *qnum; | |
1049 | QString *qstr; | |
1050 | QObject *qobj; | |
1051 | ||
1052 | /* bool */ | |
1053 | ||
6ce80fd8 | 1054 | qbool = qobject_to(QBool, qobject_from_jsonf_nofail("%i", false)); |
dfad9ec4 | 1055 | g_assert(qbool); |
fc48ffc3 | 1056 | g_assert(qbool_get_bool(qbool) == false); |
cb3e7f08 | 1057 | qobject_unref(qbool); |
e549e716 | 1058 | |
fc48ffc3 | 1059 | /* Test that non-zero values other than 1 get collapsed to true */ |
6ce80fd8 | 1060 | qbool = qobject_to(QBool, qobject_from_jsonf_nofail("%i", 2)); |
dfad9ec4 | 1061 | g_assert(qbool); |
fc48ffc3 | 1062 | g_assert(qbool_get_bool(qbool) == true); |
cb3e7f08 | 1063 | qobject_unref(qbool); |
e549e716 | 1064 | |
2e933f57 | 1065 | /* number */ |
e549e716 | 1066 | |
2e933f57 MA |
1067 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%d", value_d)); |
1068 | g_assert_cmpint(qnum_get_int(qnum), ==, value_d); | |
1069 | qobject_unref(qnum); | |
e549e716 | 1070 | |
2e933f57 MA |
1071 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%ld", value_ld)); |
1072 | g_assert_cmpint(qnum_get_int(qnum), ==, value_ld); | |
1073 | qobject_unref(qnum); | |
1074 | ||
1075 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%lld", value_lld)); | |
1076 | g_assert_cmpint(qnum_get_int(qnum), ==, value_lld); | |
1077 | qobject_unref(qnum); | |
1078 | ||
1079 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%u", value_u)); | |
1080 | g_assert_cmpuint(qnum_get_uint(qnum), ==, value_u); | |
1081 | qobject_unref(qnum); | |
1082 | ||
1083 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%lu", value_lu)); | |
1084 | g_assert_cmpuint(qnum_get_uint(qnum), ==, value_lu); | |
1085 | qobject_unref(qnum); | |
1086 | ||
1087 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%llu", value_llu)); | |
1088 | g_assert_cmpuint(qnum_get_uint(qnum), ==, value_llu); | |
1089 | qobject_unref(qnum); | |
1090 | ||
1091 | qnum = qobject_to(QNum, qobject_from_jsonf_nofail("%f", value_f)); | |
1092 | g_assert(qnum_get_double(qnum) == value_f); | |
1093 | qobject_unref(qnum); | |
1094 | ||
1095 | /* string */ | |
1096 | ||
1097 | qstr = qobject_to(QString, | |
1098 | qobject_from_jsonf_nofail("%s", value_s)); | |
1099 | g_assert_cmpstr(qstring_get_try_str(qstr), ==, value_s); | |
1100 | qobject_unref(qstr); | |
1101 | ||
1102 | /* object */ | |
1103 | ||
1104 | qobj = qobject_from_jsonf_nofail("%p", value_p); | |
1105 | g_assert(qobj == value_p); | |
1106 | } | |
1107 | ||
1108 | static void interpolation_unknown(void) | |
1109 | { | |
1110 | if (g_test_subprocess()) { | |
1111 | qobject_from_jsonf_nofail("%x", 666); | |
1112 | } | |
1113 | g_test_trap_subprocess(NULL, 0, 0); | |
1114 | g_test_trap_assert_failed(); | |
1115 | } | |
1116 | ||
1117 | static void interpolation_string(void) | |
1118 | { | |
1119 | QLitObject decoded = QLIT_QLIST(((QLitObject[]){ | |
1120 | QLIT_QSTR("%s"), | |
1121 | QLIT_QSTR("eins"), | |
1122 | {}})); | |
1123 | QObject *qobj; | |
1124 | ||
1125 | /* Dangerous misfeature: % is silently ignored in strings */ | |
1126 | qobj = qobject_from_jsonf_nofail("['%s', %s]", "eins", "zwei"); | |
1127 | g_assert(qlit_equal_qobject(&decoded, qobj)); | |
1128 | qobject_unref(qobj); | |
422c46a8 | 1129 | } |
422c46a8 | 1130 | |
ef76dc59 | 1131 | static void simple_dict(void) |
422c46a8 AL |
1132 | { |
1133 | int i; | |
1134 | struct { | |
1135 | const char *encoded; | |
082696e7 | 1136 | QLitObject decoded; |
422c46a8 AL |
1137 | } test_cases[] = { |
1138 | { | |
6ee59202 | 1139 | .encoded = "{\"foo\": 42, \"bar\": \"hello world\"}", |
082696e7 | 1140 | .decoded = QLIT_QDICT(((QLitDictEntry[]){ |
01b2ffce | 1141 | { "foo", QLIT_QNUM(42) }, |
422c46a8 AL |
1142 | { "bar", QLIT_QSTR("hello world") }, |
1143 | { } | |
1144 | })), | |
1145 | }, { | |
1146 | .encoded = "{}", | |
082696e7 | 1147 | .decoded = QLIT_QDICT(((QLitDictEntry[]){ |
422c46a8 AL |
1148 | { } |
1149 | })), | |
1150 | }, { | |
6ee59202 | 1151 | .encoded = "{\"foo\": 43}", |
082696e7 | 1152 | .decoded = QLIT_QDICT(((QLitDictEntry[]){ |
01b2ffce | 1153 | { "foo", QLIT_QNUM(43) }, |
422c46a8 AL |
1154 | { } |
1155 | })), | |
1156 | }, | |
1157 | { } | |
1158 | }; | |
1159 | ||
1160 | for (i = 0; test_cases[i].encoded; i++) { | |
1161 | QObject *obj; | |
6ee59202 | 1162 | QString *str; |
422c46a8 | 1163 | |
aec4b054 | 1164 | obj = qobject_from_json(test_cases[i].encoded, &error_abort); |
d9eba57a | 1165 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
422c46a8 | 1166 | |
6ee59202 | 1167 | str = qobject_to_json(obj); |
cb3e7f08 | 1168 | qobject_unref(obj); |
6ee59202 | 1169 | |
aec4b054 | 1170 | obj = qobject_from_json(qstring_get_str(str), &error_abort); |
d9eba57a | 1171 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
cb3e7f08 MAL |
1172 | qobject_unref(obj); |
1173 | qobject_unref(str); | |
422c46a8 AL |
1174 | } |
1175 | } | |
422c46a8 | 1176 | |
7109edfe MR |
1177 | /* |
1178 | * this generates json of the form: | |
1179 | * a(0,m) = [0, 1, ..., m-1] | |
1180 | * a(n,m) = { | |
1181 | * 'key0': a(0,m), | |
1182 | * 'key1': a(1,m), | |
1183 | * ... | |
1184 | * 'key(n-1)': a(n-1,m) | |
1185 | * } | |
1186 | */ | |
1187 | static void gen_test_json(GString *gstr, int nest_level_max, | |
1188 | int elem_count) | |
1189 | { | |
1190 | int i; | |
1191 | ||
1192 | g_assert(gstr); | |
1193 | if (nest_level_max == 0) { | |
1194 | g_string_append(gstr, "["); | |
1195 | for (i = 0; i < elem_count; i++) { | |
1196 | g_string_append_printf(gstr, "%d", i); | |
1197 | if (i < elem_count - 1) { | |
1198 | g_string_append_printf(gstr, ", "); | |
1199 | } | |
1200 | } | |
1201 | g_string_append(gstr, "]"); | |
1202 | return; | |
1203 | } | |
1204 | ||
1205 | g_string_append(gstr, "{"); | |
1206 | for (i = 0; i < nest_level_max; i++) { | |
1207 | g_string_append_printf(gstr, "'key%d': ", i); | |
1208 | gen_test_json(gstr, i, elem_count); | |
1209 | if (i < nest_level_max - 1) { | |
1210 | g_string_append(gstr, ","); | |
1211 | } | |
1212 | } | |
1213 | g_string_append(gstr, "}"); | |
1214 | } | |
1215 | ||
1216 | static void large_dict(void) | |
1217 | { | |
1218 | GString *gstr = g_string_new(""); | |
1219 | QObject *obj; | |
1220 | ||
1221 | gen_test_json(gstr, 10, 100); | |
aec4b054 | 1222 | obj = qobject_from_json(gstr->str, &error_abort); |
7109edfe MR |
1223 | g_assert(obj != NULL); |
1224 | ||
cb3e7f08 | 1225 | qobject_unref(obj); |
7109edfe MR |
1226 | g_string_free(gstr, true); |
1227 | } | |
1228 | ||
ef76dc59 | 1229 | static void simple_list(void) |
422c46a8 AL |
1230 | { |
1231 | int i; | |
1232 | struct { | |
1233 | const char *encoded; | |
082696e7 | 1234 | QLitObject decoded; |
422c46a8 AL |
1235 | } test_cases[] = { |
1236 | { | |
1237 | .encoded = "[43,42]", | |
082696e7 | 1238 | .decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce MAL |
1239 | QLIT_QNUM(43), |
1240 | QLIT_QNUM(42), | |
422c46a8 AL |
1241 | { } |
1242 | })), | |
1243 | }, | |
1244 | { | |
1245 | .encoded = "[43]", | |
082696e7 | 1246 | .decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce | 1247 | QLIT_QNUM(43), |
422c46a8 AL |
1248 | { } |
1249 | })), | |
1250 | }, | |
1251 | { | |
1252 | .encoded = "[]", | |
082696e7 | 1253 | .decoded = QLIT_QLIST(((QLitObject[]){ |
422c46a8 AL |
1254 | { } |
1255 | })), | |
1256 | }, | |
1257 | { | |
1258 | .encoded = "[{}]", | |
082696e7 MAL |
1259 | .decoded = QLIT_QLIST(((QLitObject[]){ |
1260 | QLIT_QDICT(((QLitDictEntry[]){ | |
422c46a8 AL |
1261 | {}, |
1262 | })), | |
1263 | {}, | |
1264 | })), | |
1265 | }, | |
1266 | { } | |
1267 | }; | |
1268 | ||
1269 | for (i = 0; test_cases[i].encoded; i++) { | |
1270 | QObject *obj; | |
6ee59202 | 1271 | QString *str; |
422c46a8 | 1272 | |
aec4b054 | 1273 | obj = qobject_from_json(test_cases[i].encoded, &error_abort); |
d9eba57a | 1274 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
422c46a8 | 1275 | |
6ee59202 | 1276 | str = qobject_to_json(obj); |
cb3e7f08 | 1277 | qobject_unref(obj); |
6ee59202 | 1278 | |
aec4b054 | 1279 | obj = qobject_from_json(qstring_get_str(str), &error_abort); |
d9eba57a | 1280 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
cb3e7f08 MAL |
1281 | qobject_unref(obj); |
1282 | qobject_unref(str); | |
422c46a8 AL |
1283 | } |
1284 | } | |
422c46a8 | 1285 | |
ef76dc59 | 1286 | static void simple_whitespace(void) |
422c46a8 AL |
1287 | { |
1288 | int i; | |
1289 | struct { | |
1290 | const char *encoded; | |
082696e7 | 1291 | QLitObject decoded; |
422c46a8 AL |
1292 | } test_cases[] = { |
1293 | { | |
1294 | .encoded = " [ 43 , 42 ]", | |
082696e7 | 1295 | .decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce MAL |
1296 | QLIT_QNUM(43), |
1297 | QLIT_QNUM(42), | |
422c46a8 AL |
1298 | { } |
1299 | })), | |
1300 | }, | |
1301 | { | |
53654908 | 1302 | .encoded = "\t[ 43 , { 'h' : 'b' },\r\n\t[ ], 42 ]\n", |
082696e7 | 1303 | .decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce | 1304 | QLIT_QNUM(43), |
082696e7 | 1305 | QLIT_QDICT(((QLitDictEntry[]){ |
422c46a8 AL |
1306 | { "h", QLIT_QSTR("b") }, |
1307 | { }})), | |
082696e7 | 1308 | QLIT_QLIST(((QLitObject[]){ |
422c46a8 | 1309 | { }})), |
01b2ffce | 1310 | QLIT_QNUM(42), |
422c46a8 AL |
1311 | { } |
1312 | })), | |
1313 | }, | |
1314 | { | |
1315 | .encoded = " [ 43 , { 'h' : 'b' , 'a' : 32 }, [ ], 42 ]", | |
082696e7 | 1316 | .decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce | 1317 | QLIT_QNUM(43), |
082696e7 | 1318 | QLIT_QDICT(((QLitDictEntry[]){ |
422c46a8 | 1319 | { "h", QLIT_QSTR("b") }, |
01b2ffce | 1320 | { "a", QLIT_QNUM(32) }, |
422c46a8 | 1321 | { }})), |
082696e7 | 1322 | QLIT_QLIST(((QLitObject[]){ |
422c46a8 | 1323 | { }})), |
01b2ffce | 1324 | QLIT_QNUM(42), |
422c46a8 AL |
1325 | { } |
1326 | })), | |
1327 | }, | |
1328 | { } | |
1329 | }; | |
1330 | ||
1331 | for (i = 0; test_cases[i].encoded; i++) { | |
1332 | QObject *obj; | |
6ee59202 | 1333 | QString *str; |
422c46a8 | 1334 | |
aec4b054 | 1335 | obj = qobject_from_json(test_cases[i].encoded, &error_abort); |
d9eba57a | 1336 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
422c46a8 | 1337 | |
6ee59202 | 1338 | str = qobject_to_json(obj); |
cb3e7f08 | 1339 | qobject_unref(obj); |
6ee59202 | 1340 | |
aec4b054 | 1341 | obj = qobject_from_json(qstring_get_str(str), &error_abort); |
d9eba57a | 1342 | g_assert(qlit_equal_qobject(&test_cases[i].decoded, obj)); |
6ee59202 | 1343 | |
cb3e7f08 MAL |
1344 | qobject_unref(obj); |
1345 | qobject_unref(str); | |
422c46a8 AL |
1346 | } |
1347 | } | |
422c46a8 | 1348 | |
2e933f57 | 1349 | static void simple_interpolation(void) |
422c46a8 AL |
1350 | { |
1351 | QObject *embedded_obj; | |
1352 | QObject *obj; | |
082696e7 | 1353 | QLitObject decoded = QLIT_QLIST(((QLitObject[]){ |
01b2ffce MAL |
1354 | QLIT_QNUM(1), |
1355 | QLIT_QNUM(2), | |
082696e7 | 1356 | QLIT_QLIST(((QLitObject[]){ |
01b2ffce MAL |
1357 | QLIT_QNUM(32), |
1358 | QLIT_QNUM(42), | |
422c46a8 AL |
1359 | {}})), |
1360 | {}})); | |
1361 | ||
aec4b054 | 1362 | embedded_obj = qobject_from_json("[32, 42]", &error_abort); |
ef76dc59 | 1363 | g_assert(embedded_obj != NULL); |
422c46a8 | 1364 | |
6ce80fd8 | 1365 | obj = qobject_from_jsonf_nofail("[%d, 2, %p]", 1, embedded_obj); |
d9eba57a | 1366 | g_assert(qlit_equal_qobject(&decoded, obj)); |
422c46a8 | 1367 | |
cb3e7f08 | 1368 | qobject_unref(obj); |
422c46a8 | 1369 | } |
422c46a8 | 1370 | |
ef76dc59 | 1371 | static void empty_input(void) |
7f8fca7c | 1372 | { |
a3694181 MA |
1373 | QObject *obj = qobject_from_json("", &error_abort); |
1374 | g_assert(obj == NULL); | |
1375 | } | |
1376 | ||
1377 | static void blank_input(void) | |
1378 | { | |
1379 | QObject *obj = qobject_from_json("\n ", &error_abort); | |
1380 | g_assert(obj == NULL); | |
1381 | } | |
1382 | ||
1383 | static void junk_input(void) | |
1384 | { | |
1385 | /* Note: junk within strings is covered elsewhere */ | |
1386 | Error *err = NULL; | |
1387 | QObject *obj; | |
1388 | ||
1389 | obj = qobject_from_json("@", &err); | |
1390 | g_assert(!err); /* BUG */ | |
1391 | g_assert(obj == NULL); | |
1392 | ||
6bc93a34 MA |
1393 | obj = qobject_from_json("{\x01", &err); |
1394 | g_assert(!err); /* BUG */ | |
1395 | g_assert(obj == NULL); | |
1396 | ||
a3694181 MA |
1397 | obj = qobject_from_json("[0\xFF]", &err); |
1398 | error_free_or_abort(&err); | |
1399 | g_assert(obj == NULL); | |
1400 | ||
1401 | obj = qobject_from_json("00", &err); | |
1402 | g_assert(!err); /* BUG */ | |
1403 | g_assert(obj == NULL); | |
1404 | ||
1405 | obj = qobject_from_json("[1e", &err); | |
1406 | g_assert(!err); /* BUG */ | |
1407 | g_assert(obj == NULL); | |
1408 | ||
1409 | obj = qobject_from_json("truer", &err); | |
1410 | error_free_or_abort(&err); | |
ef76dc59 | 1411 | g_assert(obj == NULL); |
7f8fca7c | 1412 | } |
7f8fca7c | 1413 | |
ef76dc59 | 1414 | static void unterminated_string(void) |
7f8fca7c | 1415 | { |
aec4b054 MA |
1416 | Error *err = NULL; |
1417 | QObject *obj = qobject_from_json("\"abc", &err); | |
1418 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1419 | g_assert(obj == NULL); |
7f8fca7c | 1420 | } |
7f8fca7c | 1421 | |
ef76dc59 | 1422 | static void unterminated_sq_string(void) |
7f8fca7c | 1423 | { |
aec4b054 MA |
1424 | Error *err = NULL; |
1425 | QObject *obj = qobject_from_json("'abc", &err); | |
1426 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1427 | g_assert(obj == NULL); |
7f8fca7c | 1428 | } |
7f8fca7c | 1429 | |
ef76dc59 | 1430 | static void unterminated_escape(void) |
7f8fca7c | 1431 | { |
aec4b054 MA |
1432 | Error *err = NULL; |
1433 | QObject *obj = qobject_from_json("\"abc\\\"", &err); | |
1434 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1435 | g_assert(obj == NULL); |
7f8fca7c | 1436 | } |
7f8fca7c | 1437 | |
ef76dc59 | 1438 | static void unterminated_array(void) |
7f8fca7c | 1439 | { |
aec4b054 MA |
1440 | Error *err = NULL; |
1441 | QObject *obj = qobject_from_json("[32", &err); | |
1442 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1443 | g_assert(obj == NULL); |
7f8fca7c | 1444 | } |
7f8fca7c | 1445 | |
ef76dc59 | 1446 | static void unterminated_array_comma(void) |
7f8fca7c | 1447 | { |
aec4b054 MA |
1448 | Error *err = NULL; |
1449 | QObject *obj = qobject_from_json("[32,", &err); | |
1450 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1451 | g_assert(obj == NULL); |
7f8fca7c | 1452 | } |
7f8fca7c | 1453 | |
ef76dc59 | 1454 | static void invalid_array_comma(void) |
7f8fca7c | 1455 | { |
aec4b054 MA |
1456 | Error *err = NULL; |
1457 | QObject *obj = qobject_from_json("[32,}", &err); | |
1458 | error_free_or_abort(&err); | |
ef76dc59 | 1459 | g_assert(obj == NULL); |
7f8fca7c | 1460 | } |
7f8fca7c | 1461 | |
ef76dc59 | 1462 | static void unterminated_dict(void) |
7f8fca7c | 1463 | { |
aec4b054 MA |
1464 | Error *err = NULL; |
1465 | QObject *obj = qobject_from_json("{'abc':32", &err); | |
1466 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1467 | g_assert(obj == NULL); |
7f8fca7c | 1468 | } |
7f8fca7c | 1469 | |
ef76dc59 | 1470 | static void unterminated_dict_comma(void) |
7f8fca7c | 1471 | { |
aec4b054 MA |
1472 | Error *err = NULL; |
1473 | QObject *obj = qobject_from_json("{'abc':32,", &err); | |
1474 | g_assert(!err); /* BUG */ | |
ef76dc59 | 1475 | g_assert(obj == NULL); |
7f8fca7c | 1476 | } |
7f8fca7c | 1477 | |
ef76dc59 | 1478 | static void invalid_dict_comma(void) |
7f8fca7c | 1479 | { |
aec4b054 MA |
1480 | Error *err = NULL; |
1481 | QObject *obj = qobject_from_json("{'abc':32,}", &err); | |
1482 | error_free_or_abort(&err); | |
ef76dc59 | 1483 | g_assert(obj == NULL); |
7f8fca7c | 1484 | } |
7f8fca7c | 1485 | |
ef76dc59 | 1486 | static void unterminated_literal(void) |
7f8fca7c | 1487 | { |
aec4b054 MA |
1488 | Error *err = NULL; |
1489 | QObject *obj = qobject_from_json("nul", &err); | |
1490 | error_free_or_abort(&err); | |
ef76dc59 | 1491 | g_assert(obj == NULL); |
7f8fca7c | 1492 | } |
7f8fca7c | 1493 | |
f0ae0304 MA |
1494 | static char *make_nest(char *buf, size_t cnt) |
1495 | { | |
1496 | memset(buf, '[', cnt - 1); | |
1497 | buf[cnt - 1] = '{'; | |
1498 | buf[cnt] = '}'; | |
1499 | memset(buf + cnt + 1, ']', cnt - 1); | |
1500 | buf[2 * cnt] = 0; | |
1501 | return buf; | |
1502 | } | |
1503 | ||
1504 | static void limits_nesting(void) | |
1505 | { | |
aec4b054 | 1506 | Error *err = NULL; |
f0ae0304 MA |
1507 | enum { max_nesting = 1024 }; /* see qobject/json-streamer.c */ |
1508 | char buf[2 * (max_nesting + 1) + 1]; | |
1509 | QObject *obj; | |
1510 | ||
aec4b054 | 1511 | obj = qobject_from_json(make_nest(buf, max_nesting), &error_abort); |
f0ae0304 | 1512 | g_assert(obj != NULL); |
cb3e7f08 | 1513 | qobject_unref(obj); |
f0ae0304 | 1514 | |
aec4b054 MA |
1515 | obj = qobject_from_json(make_nest(buf, max_nesting + 1), &err); |
1516 | error_free_or_abort(&err); | |
f0ae0304 MA |
1517 | g_assert(obj == NULL); |
1518 | } | |
1519 | ||
956a104a MA |
1520 | static void multiple_values(void) |
1521 | { | |
1522 | Error *err = NULL; | |
1523 | QObject *obj; | |
1524 | ||
1525 | /* BUG this leaks the syntax tree for "false" */ | |
1526 | obj = qobject_from_json("false true", &err); | |
1527 | g_assert(qbool_get_bool(qobject_to(QBool, obj))); | |
1528 | g_assert(!err); | |
1529 | qobject_unref(obj); | |
1530 | ||
1531 | /* BUG simultaneously succeeds and fails */ | |
1532 | /* BUG calls json_parser_parse() with errp pointing to non-null */ | |
1533 | obj = qobject_from_json("} true", &err); | |
1534 | g_assert(qbool_get_bool(qobject_to(QBool, obj))); | |
1535 | error_free_or_abort(&err); | |
1536 | qobject_unref(obj); | |
1537 | } | |
1538 | ||
ef76dc59 | 1539 | int main(int argc, char **argv) |
422c46a8 | 1540 | { |
ef76dc59 AL |
1541 | g_test_init(&argc, &argv, NULL); |
1542 | ||
ef76dc59 | 1543 | g_test_add_func("/literals/string/escaped", escaped_string); |
069946f4 | 1544 | g_test_add_func("/literals/string/quotes", string_with_quotes); |
3960c41f | 1545 | g_test_add_func("/literals/string/utf8", utf8_string); |
ef76dc59 AL |
1546 | |
1547 | g_test_add_func("/literals/number/simple", simple_number); | |
2bc7cfea | 1548 | g_test_add_func("/literals/number/large", large_number); |
ef76dc59 | 1549 | g_test_add_func("/literals/number/float", float_number); |
ef76dc59 AL |
1550 | |
1551 | g_test_add_func("/literals/keyword", keyword_literal); | |
1552 | ||
2e933f57 MA |
1553 | g_test_add_func("/literals/interpolation/valid", interpolation_valid); |
1554 | g_test_add_func("/literals/interpolation/unkown", interpolation_unknown); | |
1555 | g_test_add_func("/literals/interpolation/string", interpolation_string); | |
1556 | ||
ef76dc59 | 1557 | g_test_add_func("/dicts/simple_dict", simple_dict); |
7109edfe | 1558 | g_test_add_func("/dicts/large_dict", large_dict); |
ef76dc59 AL |
1559 | g_test_add_func("/lists/simple_list", simple_list); |
1560 | ||
2e933f57 MA |
1561 | g_test_add_func("/mixed/simple_whitespace", simple_whitespace); |
1562 | g_test_add_func("/mixed/interpolation", simple_interpolation); | |
ef76dc59 | 1563 | |
a3694181 MA |
1564 | g_test_add_func("/errors/empty", empty_input); |
1565 | g_test_add_func("/errors/blank", blank_input); | |
1566 | g_test_add_func("/errors/junk", junk_input); | |
ef76dc59 AL |
1567 | g_test_add_func("/errors/unterminated/string", unterminated_string); |
1568 | g_test_add_func("/errors/unterminated/escape", unterminated_escape); | |
1569 | g_test_add_func("/errors/unterminated/sq_string", unterminated_sq_string); | |
1570 | g_test_add_func("/errors/unterminated/array", unterminated_array); | |
1571 | g_test_add_func("/errors/unterminated/array_comma", unterminated_array_comma); | |
1572 | g_test_add_func("/errors/unterminated/dict", unterminated_dict); | |
1573 | g_test_add_func("/errors/unterminated/dict_comma", unterminated_dict_comma); | |
1574 | g_test_add_func("/errors/invalid_array_comma", invalid_array_comma); | |
ef76dc59 AL |
1575 | g_test_add_func("/errors/invalid_dict_comma", invalid_dict_comma); |
1576 | g_test_add_func("/errors/unterminated/literal", unterminated_literal); | |
f0ae0304 | 1577 | g_test_add_func("/errors/limits/nesting", limits_nesting); |
956a104a | 1578 | g_test_add_func("/errors/multiple_values", multiple_values); |
7f8fca7c | 1579 | |
ef76dc59 | 1580 | return g_test_run(); |
422c46a8 | 1581 | } |