1 /* Test of quotearg family of functions.
2 Copyright (C) 2008-2021 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 <stdbool.h>
25 #include <stdint.h>
26 #include <stdlib.h>
27 #include <string.h>
28
29 #include "localcharset.h"
30 #include "macros.h"
31 #include "zerosize-ptr.h"
32
33 #include "test-quotearg.h"
34
35 static struct result_groups results_g[] = {
36 /* literal_quoting_style */
37 { { "", "\0""1\0", 3, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
38 "a' b", LQ RQ, LQ RQ },
39 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
40 "a' b", LQ RQ, LQ RQ },
41 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b",
42 "a' b", LQ RQ, LQ RQ } },
43
44 /* shell_quoting_style */
45 { { "''", "\0""1\0", 3, "simple", "' \t\n'\\''\"\033?""?/\\'", "a:b",
46 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ },
47 { "''", "1", 1, "simple", "' \t\n'\\''\"\033?""?/\\'", "a:b",
48 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ },
49 { "''", "1", 1, "simple", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
50 "'a\\b'", "\"a' b\"", LQ RQ, LQ RQ } },
51
52 /* shell_always_quoting_style */
53 { { "''", "'\0""1\0'", 5, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
54 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" },
55 { "''", "'1'", 3, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
56 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" },
57 { "''", "'1'", 3, "'simple'", "' \t\n'\\''\"\033?""?/\\'", "'a:b'",
58 "'a\\b'", "\"a' b\"", "'" LQ RQ "'", "'" LQ RQ "'" } },
59
60 /* shell_escape_quoting_style */
61 { { "''", "''$'\\0''1'$'\\0'", 15, "simple",
62 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "a:b",
63 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ },
64 { "''", "''$'\\0''1'$'\\0'", 15, "simple",
65 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "a:b",
66 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ },
67 { "''", "''$'\\0''1'$'\\0'", 15, "simple",
68 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
69 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", LQ RQ } },
70
71 /* shell_escape_always_quoting_style */
72 { { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
73 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
74 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
75 { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
76 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
77 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
78 { "''", "''$'\\0''1'$'\\0'", 15, "'simple'",
79 "' '$'\\t\\n'\\''\"'$'\\033''?""?/\\'", "'a:b'",
80 "'a\\b'", "\"a' b\"", "''$'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
81
82 /* c_quoting_style */
83 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
84 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
85 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
86 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
87 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
88 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
89 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
90 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
91 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } },
92
93 /* c_maybe_quoting_style */
94 { { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
95 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
96 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
97 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
98 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
99 "\"a:b\"", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ } },
100
101 /* escape_quoting_style */
102 { { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a:b",
103 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ },
104 { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a:b",
105 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ },
106 { "", "\\0001\\0", 7, "simple", " \\t\\n'\"\\033?""?/\\\\", "a\\:b",
107 "a\\\\b", "a' b", LQ_ENC RQ_ENC, LQ RQ } },
108
109 /* locale_quoting_style */
110 { { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
111 "'a:b'", "'a\\\\b'", "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
112 { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
113 "'a:b'", "'a\\\\b'", "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
114 { "''", "'\\0001\\0'", 9, "'simple'", "' \\t\\n\\'\"\\033?""?/\\\\'",
115 "'a\\:b'", "'a\\\\b'", "'a\\' b'",
116 "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
117
118 /* clocale_quoting_style */
119 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
120 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
121 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
122 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
123 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
124 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
125 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
126 "\" \\t\\n'\\\"\\033?""?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
127 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } }
128 };
129
130 static struct result_groups flag_results[] = {
131 /* literal_quoting_style and QA_ELIDE_NULL_BYTES */
132 { { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
133 LQ RQ, LQ RQ },
134 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
135 LQ RQ, LQ RQ },
136 { "", "1", 1, "simple", " \t\n'\"\033?""?/\\", "a:b", "a\\b", "a' b",
137 LQ RQ, LQ RQ } },
138
139 /* c_quoting_style and QA_ELIDE_OUTER_QUOTES */
140 { { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
141 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
142 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
143 "a:b", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ },
144 { "", "\"\\0001\\0\"", 9, "simple", "\" \\t\\n'\\\"\\033?""?/\\\\\"",
145 "\"a:b\"", "a\\b", "a' b", "\"" LQ_ENC RQ_ENC "\"", LQ RQ } },
146
147 /* c_quoting_style and QA_SPLIT_TRIGRAPHS */
148 { { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
149 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
150 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
151 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
152 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a:b\"", "\"a\\\\b\"",
153 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" },
154 { "\"\"", "\"\\0001\\0\"", 9, "\"simple\"",
155 "\" \\t\\n'\\\"\\033?\"\"?/\\\\\"", "\"a\\:b\"", "\"a\\\\b\"",
156 "\"a' b\"", "\"" LQ_ENC RQ_ENC "\"", "\"" LQ RQ "\"" } }
157 };
158
159 static char const *custom_quotes[][2] = {
160 { "", "" },
161 { "'", "'" },
162 { "(", ")" },
163 { ":", " " },
164 { " ", ":" },
165 { "# ", "\n" },
166 { "\"'", "'\"" }
167 };
168
169 static struct result_groups custom_results[] = {
170 /* left_quote = right_quote = "" */
171 { { "", "\\0001\\0", 7, "simple",
172 " \\t\\n'\"\\033?""?/\\\\", "a:b", "a\\\\b",
173 "a' b", LQ_ENC RQ_ENC, LQ RQ },
174 { "", "\\0001\\0", 7, "simple",
175 " \\t\\n'\"\\033?""?/\\\\", "a:b", "a\\\\b",
176 "a' b", LQ_ENC RQ_ENC, LQ RQ },
177 { "", "\\0001\\0", 7, "simple",
178 " \\t\\n'\"\\033?""?/\\\\", "a\\:b", "a\\\\b",
179 "a' b", LQ_ENC RQ_ENC, LQ RQ } },
180
181 /* left_quote = right_quote = "'" */
182 { { "''", "'\\0001\\0'", 9, "'simple'",
183 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a:b'", "'a\\\\b'",
184 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
185 { "''", "'\\0001\\0'", 9, "'simple'",
186 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a:b'", "'a\\\\b'",
187 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" },
188 { "''", "'\\0001\\0'", 9, "'simple'",
189 "' \\t\\n\\'\"\\033?""?/\\\\'", "'a\\:b'", "'a\\\\b'",
190 "'a\\' b'", "'" LQ_ENC RQ_ENC "'", "'" LQ RQ "'" } },
191
192 /* left_quote = "(" and right_quote = ")" */
193 { { "()", "(\\0001\\0)", 9, "(simple)",
194 "( \\t\\n'\"\\033?""?/\\\\)", "(a:b)", "(a\\\\b)",
195 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" },
196 { "()", "(\\0001\\0)", 9, "(simple)",
197 "( \\t\\n'\"\\033?""?/\\\\)", "(a:b)", "(a\\\\b)",
198 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" },
199 { "()", "(\\0001\\0)", 9, "(simple)",
200 "( \\t\\n'\"\\033?""?/\\\\)", "(a\\:b)", "(a\\\\b)",
201 "(a' b)", "(" LQ_ENC RQ_ENC ")", "(" LQ RQ ")" } },
202
203 /* left_quote = ":" and right_quote = " " */
204 { { ": ", ":\\0001\\0 ", 9, ":simple ",
205 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a:b ", ":a\\\\b ",
206 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " },
207 { ": ", ":\\0001\\0 ", 9, ":simple ",
208 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a:b ", ":a\\\\b ",
209 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " },
210 { ": ", ":\\0001\\0 ", 9, ":simple ",
211 ":\\ \\t\\n'\"\\033?""?/\\\\ ", ":a\\:b ", ":a\\\\b ",
212 ":a'\\ b ", ":" LQ_ENC RQ_ENC " ", ":" LQ RQ " " } },
213
214 /* left_quote = " " and right_quote = ":" */
215 { { " :", " \\0001\\0:", 9, " simple:",
216 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
217 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" },
218 { " :", " \\0001\\0:", 9, " simple:",
219 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
220 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" },
221 { " :", " \\0001\\0:", 9, " simple:",
222 " \\t\\n'\"\\033?""?/\\\\:", " a\\:b:", " a\\\\b:",
223 " a' b:", " " LQ_ENC RQ_ENC ":", " " LQ RQ ":" } },
224
225 /* left_quote = "# " and right_quote = "\n" */
226 { { "# \n", "# \\0001\\0\n", 10, "# simple\n",
227 "# \\t\\n'\"\\033?""?/\\\\\n", "# a:b\n", "# a\\\\b\n",
228 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" },
229 { "# \n", "# \\0001\\0\n", 10, "# simple\n",
230 "# \\t\\n'\"\\033?""?/\\\\\n", "# a:b\n", "# a\\\\b\n",
231 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" },
232 { "# \n", "# \\0001\\0\n", 10, "# simple\n",
233 "# \\t\\n'\"\\033?""?/\\\\\n", "# a\\:b\n", "# a\\\\b\n",
234 "# a' b\n", "# " LQ_ENC RQ_ENC "\n", "# " LQ RQ "\n" } },
235
236 /* left_quote = "\"'" and right_quote = "'\"" */
237 { { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
238 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a:b'\"", "\"'a\\\\b'\"",
239 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" },
240 { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
241 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a:b'\"", "\"'a\\\\b'\"",
242 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" },
243 { "\"''\"", "\"'\\0001\\0'\"", 11, "\"'simple'\"",
244 "\"' \\t\\n\\'\"\\033?""?/\\\\'\"", "\"'a\\:b'\"", "\"'a\\\\b'\"",
245 "\"'a' b'\"", "\"'" LQ_ENC RQ_ENC "'\"", "\"'" LQ RQ "'\"" } }
246 };
247
248 static char *
249 use_quote_double_quotes (const char *str, size_t *len)
250 {
251 char *p = *len == SIZE_MAX ? quotearg_char (str, '"')
252 : quotearg_char_mem (str, *len, '"');
253 *len = strlen (p);
254 return p;
255 }
256
257 int
258 main (int argc _GL_UNUSED, char *argv[])
259 {
260 int i;
261 bool ascii_only = MB_CUR_MAX == 1 && !isprint ((unsigned char) LQ[0]);
262
263 /* This part of the program is hard-wired to the C locale since it
264 does not call setlocale. However, according to POSIX, the use of
265 8-bit bytes in a character context in the C locale gives
266 unspecified results (that is, the C locale charset is allowed to
267 be unibyte with 8-bit bytes rejected [ASCII], unibyte with 8-bit
268 bytes being characters [often ISO-8859-1], or multibyte [often
269 UTF-8]). We assume that the latter two cases will be
270 indistinguishable in this test - that is, the LQ and RQ sequences
271 will pass through unchanged in either type of charset. So when
272 testing for quoting of str7, use the ascii_only flag to decide
273 what to expect for the 8-bit data being quoted. */
274 ASSERT (!isprint ('\033'));
275 for (i = literal_quoting_style; i <= clocale_quoting_style; i++)
276 {
277 set_quoting_style (NULL, (enum quoting_style) i);
278 if (!(i == locale_quoting_style || i == clocale_quoting_style)
279 || (strcmp (locale_charset (), "ASCII") == 0
280 || strcmp (locale_charset (), "ANSI_X3.4-1968") == 0))
281 {
282 compare_strings (use_quotearg_buffer, &results_g[i].group1,
283 ascii_only);
284 compare_strings (use_quotearg, &results_g[i].group2,
285 ascii_only);
286 if (i == c_quoting_style)
287 compare_strings (use_quote_double_quotes, &results_g[i].group2,
288 ascii_only);
289 compare_strings (use_quotearg_colon, &results_g[i].group3,
290 ascii_only);
291 }
292 }
293
294 set_quoting_style (NULL, literal_quoting_style);
295 ASSERT (set_quoting_flags (NULL, QA_ELIDE_NULL_BYTES) == 0);
296 compare_strings (use_quotearg_buffer, &flag_results[0].group1, ascii_only);
297 compare_strings (use_quotearg, &flag_results[0].group2, ascii_only);
298 compare_strings (use_quotearg_colon, &flag_results[0].group3, ascii_only);
299
300 set_quoting_style (NULL, c_quoting_style);
301 ASSERT (set_quoting_flags (NULL, QA_ELIDE_OUTER_QUOTES)
302 == QA_ELIDE_NULL_BYTES);
303 compare_strings (use_quotearg_buffer, &flag_results[1].group1, ascii_only);
304 compare_strings (use_quotearg, &flag_results[1].group2, ascii_only);
305 compare_strings (use_quote_double_quotes, &flag_results[1].group2,
306 ascii_only);
307 compare_strings (use_quotearg_colon, &flag_results[1].group3, ascii_only);
308
309 ASSERT (set_quoting_flags (NULL, QA_SPLIT_TRIGRAPHS)
310 == QA_ELIDE_OUTER_QUOTES);
311 compare_strings (use_quotearg_buffer, &flag_results[2].group1, ascii_only);
312 compare_strings (use_quotearg, &flag_results[2].group2, ascii_only);
313 compare_strings (use_quote_double_quotes, &flag_results[2].group2,
314 ascii_only);
315 compare_strings (use_quotearg_colon, &flag_results[2].group3, ascii_only);
316
317 ASSERT (set_quoting_flags (NULL, 0) == QA_SPLIT_TRIGRAPHS);
318
319 for (i = 0; i < sizeof custom_quotes / sizeof *custom_quotes; ++i)
320 {
321 set_custom_quoting (NULL,
322 custom_quotes[i][0], custom_quotes[i][1]);
323 compare_strings (use_quotearg_buffer, &custom_results[i].group1,
324 ascii_only);
325 compare_strings (use_quotearg, &custom_results[i].group2, ascii_only);
326 compare_strings (use_quotearg_colon, &custom_results[i].group3,
327 ascii_only);
328 }
329
330 {
331 /* Trigger the bug whereby quotearg_buffer would read beyond the NUL
332 that defines the end of the string being quoted. Use an input
333 string whose NUL is the last byte before an unreadable page. */
334 char *z = zerosize_ptr ();
335
336 if (z)
337 {
338 size_t q_len = 1024;
339 char *q = malloc (q_len + 1);
340 char buf[10];
341 memset (q, 'Q', q_len);
342 q[q_len] = 0;
343
344 /* Z points to the boundary between a readable/writable page
345 and one that is neither readable nor writable. Position
346 our string so its NUL is at the end of the writable one. */
347 char const *str = "____";
348 size_t s_len = strlen (str);
349 z -= s_len + 1;
350 memcpy (z, str, s_len + 1);
351
352 set_custom_quoting (NULL, q, q);
353 /* Whether this actually triggers a SEGV depends on the
354 implementation of memcmp: whether it compares only byte-at-
355 a-time, and from left to right (no SEGV) or some other way. */
356 size_t n = quotearg_buffer (buf, sizeof buf, z, SIZE_MAX, NULL);
357 ASSERT (n == s_len + 2 * q_len);
358 ASSERT (memcmp (buf, q, sizeof buf) == 0);
359 free (q);
360 }
361 }
362
363 quotearg_free ();
364
365 return 0;
366 }