]>
Commit | Line | Data |
---|---|---|
53e1b683 | 1 | /* SPDX-License-Identifier: LGPL-2.1+ */ |
4f5dd394 LP |
2 | /*** |
3 | This file is part of systemd. | |
4 | ||
5 | Copyright 2010 Lennart Poettering | |
6 | ||
7 | systemd is free software; you can redistribute it and/or modify it | |
8 | under the terms of the GNU Lesser General Public License as published by | |
9 | the Free Software Foundation; either version 2.1 of the License, or | |
10 | (at your option) any later version. | |
11 | ||
12 | systemd is distributed in the hope that it will be useful, but | |
13 | WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
15 | Lesser General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU Lesser General Public License | |
18 | along with systemd; If not, see <http://www.gnu.org/licenses/>. | |
19 | ***/ | |
20 | ||
11c3a366 TA |
21 | #include <errno.h> |
22 | #include <stdlib.h> | |
23 | #include <string.h> | |
24 | ||
b5efdb8a | 25 | #include "alloc-util.h" |
e4e73a63 LP |
26 | #include "escape.h" |
27 | #include "hexdecoct.h" | |
11c3a366 | 28 | #include "macro.h" |
4f5dd394 | 29 | #include "utf8.h" |
4f5dd394 | 30 | |
4f5dd394 LP |
31 | size_t cescape_char(char c, char *buf) { |
32 | char * buf_old = buf; | |
33 | ||
34 | switch (c) { | |
35 | ||
36 | case '\a': | |
37 | *(buf++) = '\\'; | |
38 | *(buf++) = 'a'; | |
39 | break; | |
40 | case '\b': | |
41 | *(buf++) = '\\'; | |
42 | *(buf++) = 'b'; | |
43 | break; | |
44 | case '\f': | |
45 | *(buf++) = '\\'; | |
46 | *(buf++) = 'f'; | |
47 | break; | |
48 | case '\n': | |
49 | *(buf++) = '\\'; | |
50 | *(buf++) = 'n'; | |
51 | break; | |
52 | case '\r': | |
53 | *(buf++) = '\\'; | |
54 | *(buf++) = 'r'; | |
55 | break; | |
56 | case '\t': | |
57 | *(buf++) = '\\'; | |
58 | *(buf++) = 't'; | |
59 | break; | |
60 | case '\v': | |
61 | *(buf++) = '\\'; | |
62 | *(buf++) = 'v'; | |
63 | break; | |
64 | case '\\': | |
65 | *(buf++) = '\\'; | |
66 | *(buf++) = '\\'; | |
67 | break; | |
68 | case '"': | |
69 | *(buf++) = '\\'; | |
70 | *(buf++) = '"'; | |
71 | break; | |
72 | case '\'': | |
73 | *(buf++) = '\\'; | |
74 | *(buf++) = '\''; | |
75 | break; | |
76 | ||
77 | default: | |
78 | /* For special chars we prefer octal over | |
79 | * hexadecimal encoding, simply because glib's | |
80 | * g_strescape() does the same */ | |
81 | if ((c < ' ') || (c >= 127)) { | |
82 | *(buf++) = '\\'; | |
83 | *(buf++) = octchar((unsigned char) c >> 6); | |
84 | *(buf++) = octchar((unsigned char) c >> 3); | |
85 | *(buf++) = octchar((unsigned char) c); | |
86 | } else | |
87 | *(buf++) = c; | |
88 | break; | |
89 | } | |
90 | ||
91 | return buf - buf_old; | |
92 | } | |
93 | ||
a5ef3638 | 94 | char *cescape_length(const char *s, size_t n) { |
4f5dd394 | 95 | const char *f; |
a5ef3638 | 96 | char *r, *t; |
4f5dd394 | 97 | |
a5ef3638 | 98 | assert(s || n == 0); |
4f5dd394 LP |
99 | |
100 | /* Does C style string escaping. May be reversed with | |
101 | * cunescape(). */ | |
102 | ||
a5ef3638 | 103 | r = new(char, n*4 + 1); |
4f5dd394 LP |
104 | if (!r) |
105 | return NULL; | |
106 | ||
a5ef3638 | 107 | for (f = s, t = r; f < s + n; f++) |
4f5dd394 LP |
108 | t += cescape_char(*f, t); |
109 | ||
110 | *t = 0; | |
111 | ||
112 | return r; | |
113 | } | |
114 | ||
a5ef3638 LP |
115 | char *cescape(const char *s) { |
116 | assert(s); | |
117 | ||
118 | return cescape_length(s, strlen(s)); | |
119 | } | |
120 | ||
c932fb71 | 121 | int cunescape_one(const char *p, size_t length, char32_t *ret, bool *eight_bit) { |
4f5dd394 LP |
122 | int r = 1; |
123 | ||
124 | assert(p); | |
125 | assert(*p); | |
126 | assert(ret); | |
127 | ||
3565e095 ZJS |
128 | /* Unescapes C style. Returns the unescaped character in ret. |
129 | * Sets *eight_bit to true if the escaped sequence either fits in | |
130 | * one byte in UTF-8 or is a non-unicode literal byte and should | |
131 | * instead be copied directly. | |
132 | */ | |
4f5dd394 LP |
133 | |
134 | if (length != (size_t) -1 && length < 1) | |
135 | return -EINVAL; | |
136 | ||
137 | switch (p[0]) { | |
138 | ||
139 | case 'a': | |
140 | *ret = '\a'; | |
141 | break; | |
142 | case 'b': | |
143 | *ret = '\b'; | |
144 | break; | |
145 | case 'f': | |
146 | *ret = '\f'; | |
147 | break; | |
148 | case 'n': | |
149 | *ret = '\n'; | |
150 | break; | |
151 | case 'r': | |
152 | *ret = '\r'; | |
153 | break; | |
154 | case 't': | |
155 | *ret = '\t'; | |
156 | break; | |
157 | case 'v': | |
158 | *ret = '\v'; | |
159 | break; | |
160 | case '\\': | |
161 | *ret = '\\'; | |
162 | break; | |
163 | case '"': | |
164 | *ret = '"'; | |
165 | break; | |
166 | case '\'': | |
167 | *ret = '\''; | |
168 | break; | |
169 | ||
170 | case 's': | |
171 | /* This is an extension of the XDG syntax files */ | |
172 | *ret = ' '; | |
173 | break; | |
174 | ||
175 | case 'x': { | |
176 | /* hexadecimal encoding */ | |
177 | int a, b; | |
178 | ||
179 | if (length != (size_t) -1 && length < 3) | |
180 | return -EINVAL; | |
181 | ||
182 | a = unhexchar(p[1]); | |
183 | if (a < 0) | |
184 | return -EINVAL; | |
185 | ||
186 | b = unhexchar(p[2]); | |
187 | if (b < 0) | |
188 | return -EINVAL; | |
189 | ||
190 | /* Don't allow NUL bytes */ | |
191 | if (a == 0 && b == 0) | |
192 | return -EINVAL; | |
193 | ||
3565e095 ZJS |
194 | *ret = (a << 4U) | b; |
195 | *eight_bit = true; | |
4f5dd394 LP |
196 | r = 3; |
197 | break; | |
198 | } | |
199 | ||
200 | case 'u': { | |
201 | /* C++11 style 16bit unicode */ | |
202 | ||
203 | int a[4]; | |
204 | unsigned i; | |
205 | uint32_t c; | |
206 | ||
207 | if (length != (size_t) -1 && length < 5) | |
208 | return -EINVAL; | |
209 | ||
210 | for (i = 0; i < 4; i++) { | |
211 | a[i] = unhexchar(p[1 + i]); | |
212 | if (a[i] < 0) | |
213 | return a[i]; | |
214 | } | |
215 | ||
216 | c = ((uint32_t) a[0] << 12U) | ((uint32_t) a[1] << 8U) | ((uint32_t) a[2] << 4U) | (uint32_t) a[3]; | |
217 | ||
218 | /* Don't allow 0 chars */ | |
219 | if (c == 0) | |
220 | return -EINVAL; | |
221 | ||
3565e095 | 222 | *ret = c; |
4f5dd394 LP |
223 | r = 5; |
224 | break; | |
225 | } | |
226 | ||
227 | case 'U': { | |
228 | /* C++11 style 32bit unicode */ | |
229 | ||
230 | int a[8]; | |
231 | unsigned i; | |
c932fb71 | 232 | char32_t c; |
4f5dd394 LP |
233 | |
234 | if (length != (size_t) -1 && length < 9) | |
235 | return -EINVAL; | |
236 | ||
237 | for (i = 0; i < 8; i++) { | |
238 | a[i] = unhexchar(p[1 + i]); | |
239 | if (a[i] < 0) | |
240 | return a[i]; | |
241 | } | |
242 | ||
dcd12626 LP |
243 | c = ((uint32_t) a[0] << 28U) | ((uint32_t) a[1] << 24U) | ((uint32_t) a[2] << 20U) | ((uint32_t) a[3] << 16U) | |
244 | ((uint32_t) a[4] << 12U) | ((uint32_t) a[5] << 8U) | ((uint32_t) a[6] << 4U) | (uint32_t) a[7]; | |
4f5dd394 LP |
245 | |
246 | /* Don't allow 0 chars */ | |
247 | if (c == 0) | |
248 | return -EINVAL; | |
249 | ||
250 | /* Don't allow invalid code points */ | |
251 | if (!unichar_is_valid(c)) | |
252 | return -EINVAL; | |
253 | ||
3565e095 | 254 | *ret = c; |
4f5dd394 LP |
255 | r = 9; |
256 | break; | |
257 | } | |
258 | ||
259 | case '0': | |
260 | case '1': | |
261 | case '2': | |
262 | case '3': | |
263 | case '4': | |
264 | case '5': | |
265 | case '6': | |
266 | case '7': { | |
267 | /* octal encoding */ | |
268 | int a, b, c; | |
c932fb71 | 269 | char32_t m; |
4f5dd394 LP |
270 | |
271 | if (length != (size_t) -1 && length < 3) | |
272 | return -EINVAL; | |
273 | ||
274 | a = unoctchar(p[0]); | |
275 | if (a < 0) | |
276 | return -EINVAL; | |
277 | ||
278 | b = unoctchar(p[1]); | |
279 | if (b < 0) | |
280 | return -EINVAL; | |
281 | ||
282 | c = unoctchar(p[2]); | |
283 | if (c < 0) | |
284 | return -EINVAL; | |
285 | ||
286 | /* don't allow NUL bytes */ | |
287 | if (a == 0 && b == 0 && c == 0) | |
288 | return -EINVAL; | |
289 | ||
290 | /* Don't allow bytes above 255 */ | |
dcd12626 | 291 | m = ((uint32_t) a << 6U) | ((uint32_t) b << 3U) | (uint32_t) c; |
4f5dd394 LP |
292 | if (m > 255) |
293 | return -EINVAL; | |
294 | ||
295 | *ret = m; | |
3565e095 | 296 | *eight_bit = true; |
4f5dd394 LP |
297 | r = 3; |
298 | break; | |
299 | } | |
300 | ||
301 | default: | |
302 | return -EINVAL; | |
303 | } | |
304 | ||
305 | return r; | |
306 | } | |
307 | ||
308 | int cunescape_length_with_prefix(const char *s, size_t length, const char *prefix, UnescapeFlags flags, char **ret) { | |
309 | char *r, *t; | |
310 | const char *f; | |
311 | size_t pl; | |
312 | ||
313 | assert(s); | |
314 | assert(ret); | |
315 | ||
316 | /* Undoes C style string escaping, and optionally prefixes it. */ | |
317 | ||
7bf7ce28 | 318 | pl = strlen_ptr(prefix); |
4f5dd394 LP |
319 | |
320 | r = new(char, pl+length+1); | |
321 | if (!r) | |
322 | return -ENOMEM; | |
323 | ||
324 | if (prefix) | |
325 | memcpy(r, prefix, pl); | |
326 | ||
327 | for (f = s, t = r + pl; f < s + length; f++) { | |
328 | size_t remaining; | |
3565e095 | 329 | bool eight_bit = false; |
c932fb71 | 330 | char32_t u; |
4f5dd394 LP |
331 | int k; |
332 | ||
333 | remaining = s + length - f; | |
334 | assert(remaining > 0); | |
335 | ||
336 | if (*f != '\\') { | |
629ff674 | 337 | /* A literal, copy verbatim */ |
4f5dd394 LP |
338 | *(t++) = *f; |
339 | continue; | |
340 | } | |
341 | ||
342 | if (remaining == 1) { | |
343 | if (flags & UNESCAPE_RELAX) { | |
344 | /* A trailing backslash, copy verbatim */ | |
345 | *(t++) = *f; | |
346 | continue; | |
347 | } | |
348 | ||
349 | free(r); | |
350 | return -EINVAL; | |
351 | } | |
352 | ||
3565e095 | 353 | k = cunescape_one(f + 1, remaining - 1, &u, &eight_bit); |
4f5dd394 LP |
354 | if (k < 0) { |
355 | if (flags & UNESCAPE_RELAX) { | |
356 | /* Invalid escape code, let's take it literal then */ | |
357 | *(t++) = '\\'; | |
358 | continue; | |
359 | } | |
360 | ||
361 | free(r); | |
362 | return k; | |
363 | } | |
364 | ||
3565e095 ZJS |
365 | f += k; |
366 | if (eight_bit) | |
367 | /* One byte? Set directly as specified */ | |
368 | *(t++) = u; | |
4f5dd394 | 369 | else |
3565e095 | 370 | /* Otherwise encode as multi-byte UTF-8 */ |
4f5dd394 | 371 | t += utf8_encode_unichar(t, u); |
4f5dd394 LP |
372 | } |
373 | ||
374 | *t = 0; | |
375 | ||
376 | *ret = r; | |
377 | return t - r; | |
378 | } | |
379 | ||
380 | int cunescape_length(const char *s, size_t length, UnescapeFlags flags, char **ret) { | |
381 | return cunescape_length_with_prefix(s, length, NULL, flags, ret); | |
382 | } | |
383 | ||
384 | int cunescape(const char *s, UnescapeFlags flags, char **ret) { | |
385 | return cunescape_length(s, strlen(s), flags, ret); | |
386 | } | |
387 | ||
388 | char *xescape(const char *s, const char *bad) { | |
389 | char *r, *t; | |
390 | const char *f; | |
391 | ||
392 | /* Escapes all chars in bad, in addition to \ and all special | |
393 | * chars, in \xFF style escaping. May be reversed with | |
394 | * cunescape(). */ | |
395 | ||
396 | r = new(char, strlen(s) * 4 + 1); | |
397 | if (!r) | |
398 | return NULL; | |
399 | ||
400 | for (f = s, t = r; *f; f++) { | |
401 | ||
402 | if ((*f < ' ') || (*f >= 127) || | |
403 | (*f == '\\') || strchr(bad, *f)) { | |
404 | *(t++) = '\\'; | |
405 | *(t++) = 'x'; | |
406 | *(t++) = hexchar(*f >> 4); | |
407 | *(t++) = hexchar(*f); | |
408 | } else | |
409 | *(t++) = *f; | |
410 | } | |
411 | ||
412 | *t = 0; | |
413 | ||
414 | return r; | |
415 | } | |
416 | ||
95052df3 ZJS |
417 | char *octescape(const char *s, size_t len) { |
418 | char *r, *t; | |
419 | const char *f; | |
420 | ||
421 | /* Escapes all chars in bad, in addition to \ and " chars, | |
422 | * in \nnn style escaping. */ | |
423 | ||
424 | r = new(char, len * 4 + 1); | |
425 | if (!r) | |
426 | return NULL; | |
427 | ||
428 | for (f = s, t = r; f < s + len; f++) { | |
429 | ||
4c701096 | 430 | if (*f < ' ' || *f >= 127 || IN_SET(*f, '\\', '"')) { |
95052df3 ZJS |
431 | *(t++) = '\\'; |
432 | *(t++) = '0' + (*f >> 6); | |
433 | *(t++) = '0' + ((*f >> 3) & 8); | |
434 | *(t++) = '0' + (*f & 8); | |
435 | } else | |
436 | *(t++) = *f; | |
437 | } | |
438 | ||
439 | *t = 0; | |
440 | ||
441 | return r; | |
442 | ||
443 | } | |
444 | ||
804ee07c | 445 | static char *strcpy_backslash_escaped(char *t, const char *s, const char *bad, bool escape_tab_nl) { |
4f5dd394 LP |
446 | assert(bad); |
447 | ||
448 | for (; *s; s++) { | |
804ee07c ZJS |
449 | if (escape_tab_nl && IN_SET(*s, '\n', '\t')) { |
450 | *(t++) = '\\'; | |
451 | *(t++) = *s == '\n' ? 'n' : 't'; | |
452 | continue; | |
453 | } | |
454 | ||
4f5dd394 LP |
455 | if (*s == '\\' || strchr(bad, *s)) |
456 | *(t++) = '\\'; | |
457 | ||
458 | *(t++) = *s; | |
459 | } | |
460 | ||
461 | return t; | |
462 | } | |
463 | ||
464 | char *shell_escape(const char *s, const char *bad) { | |
465 | char *r, *t; | |
466 | ||
467 | r = new(char, strlen(s)*2+1); | |
468 | if (!r) | |
469 | return NULL; | |
470 | ||
804ee07c | 471 | t = strcpy_backslash_escaped(r, s, bad, false); |
4f5dd394 LP |
472 | *t = 0; |
473 | ||
474 | return r; | |
475 | } | |
476 | ||
804ee07c | 477 | char* shell_maybe_quote(const char *s, EscapeStyle style) { |
4f5dd394 LP |
478 | const char *p; |
479 | char *r, *t; | |
480 | ||
481 | assert(s); | |
482 | ||
804ee07c ZJS |
483 | /* Encloses a string in quotes if necessary to make it OK as a shell |
484 | * string. Note that we treat benign UTF-8 characters as needing | |
485 | * escaping too, but that should be OK. */ | |
4f5dd394 LP |
486 | |
487 | for (p = s; *p; p++) | |
488 | if (*p <= ' ' || | |
489 | *p >= 127 || | |
490 | strchr(SHELL_NEED_QUOTES, *p)) | |
491 | break; | |
492 | ||
493 | if (!*p) | |
494 | return strdup(s); | |
495 | ||
804ee07c | 496 | r = new(char, (style == ESCAPE_POSIX) + 1 + strlen(s)*2 + 1 + 1); |
4f5dd394 LP |
497 | if (!r) |
498 | return NULL; | |
499 | ||
500 | t = r; | |
804ee07c ZJS |
501 | if (style == ESCAPE_BACKSLASH) |
502 | *(t++) = '"'; | |
503 | else if (style == ESCAPE_POSIX) { | |
504 | *(t++) = '$'; | |
505 | *(t++) = '\''; | |
506 | } else | |
507 | assert_not_reached("Bad EscapeStyle"); | |
508 | ||
4f5dd394 LP |
509 | t = mempcpy(t, s, p - s); |
510 | ||
804ee07c ZJS |
511 | if (style == ESCAPE_BACKSLASH) |
512 | t = strcpy_backslash_escaped(t, p, SHELL_NEED_ESCAPE, false); | |
513 | else | |
514 | t = strcpy_backslash_escaped(t, p, SHELL_NEED_ESCAPE_POSIX, true); | |
4f5dd394 | 515 | |
804ee07c ZJS |
516 | if (style == ESCAPE_BACKSLASH) |
517 | *(t++) = '"'; | |
518 | else | |
519 | *(t++) = '\''; | |
4f5dd394 LP |
520 | *t = 0; |
521 | ||
522 | return r; | |
523 | } |