1 /* Test of quotearg family of functions.
2 Copyright (C) 2008-2023 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, or (at your option)
7 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 <https://www.gnu.org/licenses/>. */
16
17 /* Written by Eric Blake <ebb9@byu.net>, 2008. */
18
19 #include <config.h>
20
21 #include "quotearg.h"
22
23 #include <ctype.h>
24 #include <stdint.h>
25 #include <stdlib.h>
26 #include <string.h>
27
28 #include "localcharset.h"
29 #include "macros.h"
30 #include "zerosize-ptr.h"
31
32 #include "test-quotearg.h"
33
34 static struct result_groups results_g[] = {
35 /* literal_quoting_style */
36 { { "", "\0""1\0", 3, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
37 "a' b", LQ RQ, LQ RQ },
38 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
39 "a' b", LQ RQ, LQ RQ },
40 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
41 "a' b", LQ RQ, LQ RQ } },
42
43 /* shell_quoting_style */
44 { { "''", "\0""1\0", 3, "simple", "' \t\n'\\''\"\033?""?/\\'", "a:b",
45 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ },
46 { "''", "1", 1, "simple", "' \t\n'\\''\"\033?""?/\\'", "a:b",
47 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ },
48 { "''", "1", 1, "simple", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
49 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ } },
50
51 /* shell_always_quoting_style */
52 { { "''", "'\0""1\0'", 5, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
53 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" },
54 { "''", "'1'", 3, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
55 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" },
56 { "''", "'1'", 3, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
57 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" } },
58
59 /* shell_escape_quoting_style */
60 { { "''", "''$'\\0''1'$'\\0'", 15, "simple",
61 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "a:b",
62 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ },
63 { "''", "''$'\\0''1'$'\\0'", 15, "simple",
64 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "a:b",
65 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ },
66 { "''", "''$'\\0''1'$'\\0'", 15, "simple",
67 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
68 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ } },
69
70 /* shell_escape_always_quoting_style */
71 { { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
72 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
73 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
74 { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
75 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
76 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
77 { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
78 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
79 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
80
81 /* c_quoting_style */
82 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
83 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
84 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
85 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
86 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
87 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
88 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
89 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
90 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } },
91
92 /* c_maybe_quoting_style */
93 { { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
94 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
95 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
96 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
97 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
98 "\"a:b\"", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ } },
99
100 /* escape_quoting_style */
101 { { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a:b",
102 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ },
103 { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a:b",
104 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ },
105 { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a\\:b",
106 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ } },
107
108 /* locale_quoting_style */
109 { { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
110 "'a:b'", "'a\\\\b'", "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
111 { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
112 "'a:b'", "'a\\\\b'", "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
113 { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
114 "'a\\:b'", "'a\\\\b'", "'a\\' b'",
115 "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
116
117 /* clocale_quoting_style */
118 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
119 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
120 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
121 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
122 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
123 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
124 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
125 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
126 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } }
127 };
128
129 static struct result_groups flag_results[] = {
130 /* literal_quoting_style and QA_ELIDE_NULL_BYTES */
131 { { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
132 LQ RQ, LQ RQ },
133 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
134 LQ RQ, LQ RQ },
135 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
136 LQ RQ, LQ RQ } },
137
138 /* c_quoting_style and QA_ELIDE_OUTER_QUOTES */
139 { { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
140 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
141 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
142 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
143 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
144 "\"a:b\"", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ } },
145
146 /* c_quoting_style and QA_SPLIT_TRIGRAPHS */
147 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
148 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
149 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
150 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
151 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
152 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
153 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
154 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
155 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } }
156 };
157
158 static char const *custom_quotes[][2] = {
159 { "", "" },
160 { "'", "'" },
161 { "(", ")" },
162 { ":", " " },
163 { " ", ":" },
164 { "# ", "\n" },
165 { "\"'", "'\"" }
166 };
167
168 static struct result_groups custom_results[] = {
169 /* left_quote = right_quote = "" */
170 { { "", "\\0001\\0", 7, "simple",
171 " \\t\\n'\"\\033?""?/\\\\", "a:b", "a\\\\b",
172 "a' b", LQ_ENC RQ_ENC, LQ RQ },
173 { "", "\\0001\\0", 7, "simple",
174 " \\t\\n'\"\\033?""?/\\\\", "a:b", "a\\\\b",
175 "a' b", LQ_ENC RQ_ENC, LQ RQ },
176 { "", "\\0001\\0", 7, "simple",
177 " \\t\\n'\"\\033?""?/\\\\", "a\\:b", "a\\\\b",
178 "a' b", LQ_ENC RQ_ENC, LQ RQ } },
179
180 /* left_quote = right_quote = "'" */
181 { { "''", "'\\0001\\0'", 9, "'simple'",
182 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a:b'", "'a\\\\b'",
183 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
184 { "''", "'\\0001\\0'", 9, "'simple'",
185 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a:b'", "'a\\\\b'",
186 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
187 { "''", "'\\0001\\0'", 9, "'simple'",
188 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a\\:b'", "'a\\\\b'",
189 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
190
191 /* left_quote = "(" and right_quote = ")" */
192 { { "()", "(\\0001\\0)", 9, "(simple)",
193 "( \\t\\n'\"\\033?""?/\\\\)", "(a:b)", "(a\\\\b)",
194 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" },
195 { "()", "(\\0001\\0)", 9, "(simple)",
196 "( \\t\\n'\"\\033?""?/\\\\)", "(a:b)", "(a\\\\b)",
197 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" },
198 { "()", "(\\0001\\0)", 9, "(simple)",
199 "( \\t\\n'\"\\033?""?/\\\\)", "(a\\:b)", "(a\\\\b)",
200 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" } },
201
202 /* left_quote = ":" and right_quote = " " */
203 { { ": ", ":\\0001\\0 ", 9, ":simple ",
204 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a:b ", ":a\\\\b ",
205 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " },
206 { ": ", ":\\0001\\0 ", 9, ":simple ",
207 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a:b ", ":a\\\\b ",
208 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " },
209 { ": ", ":\\0001\\0 ", 9, ":simple ",
210 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a\\:b ", ":a\\\\b ",
211 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " } },
212
213 /* left_quote = " " and right_quote = ":" */
214 { { " :", " \\0001\\0:", 9, " simple:",
215 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
216 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" },
217 { " :", " \\0001\\0:", 9, " simple:",
218 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
219 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" },
220 { " :", " \\0001\\0:", 9, " simple:",
221 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
222 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" } },
223
224 /* left_quote = "# " and right_quote = "\n" */
225 { { "# \n", "# \\0001\\0\n", 10, "# simple\n",
226 "# \\t\\n'\"\\033?""?/\\\\\n", "# a:b\n", "# a\\\\b\n",
227 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" },
228 { "# \n", "# \\0001\\0\n", 10, "# simple\n",
229 "# \\t\\n'\"\\033?""?/\\\\\n", "# a:b\n", "# a\\\\b\n",
230 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" },
231 { "# \n", "# \\0001\\0\n", 10, "# simple\n",
232 "# \\t\\n'\"\\033?""?/\\\\\n", "# a\\:b\n", "# a\\\\b\n",
233 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" } },
234
235 /* left_quote = "\"'" and right_quote = "'\"" */
236 { { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
237 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a:b'\"", "\"'a\\\\b'\"",
238 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" },
239 { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
240 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a:b'\"", "\"'a\\\\b'\"",
241 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" },
242 { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
243 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a\\:b'\"", "\"'a\\\\b'\"",
244 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" } }
245 };
246
247 static char *
248 use_quote_double_quotes (const char *str, size_t *len)
249 {
250 char *p = *len == SIZE_MAX ? quotearg_char (str, '"')
251 : quotearg_char_mem (str, *len, '"');
252 *len = strlen (p);
253 return p;
254 }
255
256 int
257 main (_GL_UNUSED int argc, char *argv[])
258 {
259 int i;
260 bool ascii_only = MB_CUR_MAX == 1 && !isprint ((unsigned char) LQ[0]);
261
262 /* This part of the program is hard-wired to the C locale since it
263 does not call setlocale. However, according to POSIX, the use of
264 8-bit bytes in a character context in the C locale gives
265 unspecified results (that is, the C locale charset is allowed to
266 be unibyte with 8-bit bytes rejected [ASCII], unibyte with 8-bit
267 bytes being characters [often ISO-8859-1], or multibyte [often
268 UTF-8]). We assume that the latter two cases will be
269 indistinguishable in this test - that is, the LQ and RQ sequences
270 will pass through unchanged in either type of charset. So when
271 testing for quoting of str7, use the ascii_only flag to decide
272 what to expect for the 8-bit data being quoted. */
273 ASSERT (!isprint ('\033'));
274 for (i = literal_quoting_style; i <= clocale_quoting_style; i++)
275 {
276 set_quoting_style (NULL, (enum quoting_style) i);
277 if (!(i == locale_quoting_style || i == clocale_quoting_style)
278 || (strcmp (locale_charset (), "ASCII") == 0
279 || strcmp (locale_charset (), "ANSI_X3.4-1968") == 0))
280 {
281 compare_strings (use_quotearg_buffer, &results_g[i].group1,
282 ascii_only);
283 compare_strings (use_quotearg, &results_g[i].group2,
284 ascii_only);
285 if (i == c_quoting_style)
286 compare_strings (use_quote_double_quotes, &results_g[i].group2,
287 ascii_only);
288 compare_strings (use_quotearg_colon, &results_g[i].group3,
289 ascii_only);
290 }
291 }
292
293 set_quoting_style (NULL, literal_quoting_style);
294 ASSERT (set_quoting_flags (NULL, QA_ELIDE_NULL_BYTES) == 0);
295 compare_strings (use_quotearg_buffer, &flag_results[0].group1, ascii_only);
296 compare_strings (use_quotearg, &flag_results[0].group2, ascii_only);
297 compare_strings (use_quotearg_colon, &flag_results[0].group3, ascii_only);
298
299 set_quoting_style (NULL, c_quoting_style);
300 ASSERT (set_quoting_flags (NULL, QA_ELIDE_OUTER_QUOTES)
301 == QA_ELIDE_NULL_BYTES);
302 compare_strings (use_quotearg_buffer, &flag_results[1].group1, ascii_only);
303 compare_strings (use_quotearg, &flag_results[1].group2, ascii_only);
304 compare_strings (use_quote_double_quotes, &flag_results[1].group2,
305 ascii_only);
306 compare_strings (use_quotearg_colon, &flag_results[1].group3, ascii_only);
307
308 ASSERT (set_quoting_flags (NULL, QA_SPLIT_TRIGRAPHS)
309 == QA_ELIDE_OUTER_QUOTES);
310 compare_strings (use_quotearg_buffer, &flag_results[2].group1, ascii_only);
311 compare_strings (use_quotearg, &flag_results[2].group2, ascii_only);
312 compare_strings (use_quote_double_quotes, &flag_results[2].group2,
313 ascii_only);
314 compare_strings (use_quotearg_colon, &flag_results[2].group3, ascii_only);
315
316 ASSERT (set_quoting_flags (NULL, 0) == QA_SPLIT_TRIGRAPHS);
317
318 for (i = 0; i < sizeof custom_quotes / sizeof *custom_quotes; ++i)
319 {
320 set_custom_quoting (NULL,
321 custom_quotes[i][0], custom_quotes[i][1]);
322 compare_strings (use_quotearg_buffer, &custom_results[i].group1,
323 ascii_only);
324 compare_strings (use_quotearg, &custom_results[i].group2, ascii_only);
325 compare_strings (use_quotearg_colon, &custom_results[i].group3,
326 ascii_only);
327 }
328
329 {
330 /* Trigger the bug whereby quotearg_buffer would read beyond the NUL
331 that defines the end of the string being quoted. Use an input
332 string whose NUL is the last byte before an unreadable page. */
333 char *z = zerosize_ptr ();
334
335 if (z)
336 {
337 size_t q_len = 1024;
338 char *q = malloc (q_len + 1);
339 char buf[10];
340 memset (q, 'Q', q_len);
341 q[q_len] = 0;
342
343 /* Z points to the boundary between a readable/writable page
344 and one that is neither readable nor writable. Position
345 our string so its NUL is at the end of the writable one. */
346 char const *str = "____";
347 size_t s_len = strlen (str);
348 z -= s_len + 1;
349 memcpy (z, str, s_len + 1);
350
351 set_custom_quoting (NULL, q, q);
352 /* Whether this actually triggers a SEGV depends on the
353 implementation of memcmp: whether it compares only byte-at-
354 a-time, and from left to right (no SEGV) or some other way. */
355 size_t n = quotearg_buffer (buf, sizeof buf, z, SIZE_MAX, NULL);
356 ASSERT (n == s_len + 2 * q_len);
357 ASSERT (memcmp (buf, q, sizeof buf) == 0);
358 free (q);
359 }
360 }
361
362 quotearg_free ();
363
364 return 0;
365 }