]>
git.ipfire.org Git - thirdparty/systemd.git/blob - src/basic/escape.c
1 /* SPDX-License-Identifier: LGPL-2.1+ */
3 This file is part of systemd.
5 Copyright 2010 Lennart Poettering
12 #include "alloc-util.h"
14 #include "hexdecoct.h"
18 size_t cescape_char(char c
, char *buf
) {
65 /* For special chars we prefer octal over
66 * hexadecimal encoding, simply because glib's
67 * g_strescape() does the same */
68 if ((c
< ' ') || (c
>= 127)) {
70 *(buf
++) = octchar((unsigned char) c
>> 6);
71 *(buf
++) = octchar((unsigned char) c
>> 3);
72 *(buf
++) = octchar((unsigned char) c
);
81 char *cescape_length(const char *s
, size_t n
) {
87 /* Does C style string escaping. May be reversed with
90 r
= new(char, n
*4 + 1);
94 for (f
= s
, t
= r
; f
< s
+ n
; f
++)
95 t
+= cescape_char(*f
, t
);
102 char *cescape(const char *s
) {
105 return cescape_length(s
, strlen(s
));
108 int cunescape_one(const char *p
, size_t length
, char32_t
*ret
, bool *eight_bit
) {
115 /* Unescapes C style. Returns the unescaped character in ret.
116 * Sets *eight_bit to true if the escaped sequence either fits in
117 * one byte in UTF-8 or is a non-unicode literal byte and should
118 * instead be copied directly.
121 if (length
!= (size_t) -1 && length
< 1)
158 /* This is an extension of the XDG syntax files */
163 /* hexadecimal encoding */
166 if (length
!= (size_t) -1 && length
< 3)
177 /* Don't allow NUL bytes */
178 if (a
== 0 && b
== 0)
181 *ret
= (a
<< 4U) | b
;
188 /* C++11 style 16bit unicode */
194 if (length
!= (size_t) -1 && length
< 5)
197 for (i
= 0; i
< 4; i
++) {
198 a
[i
] = unhexchar(p
[1 + i
]);
203 c
= ((uint32_t) a
[0] << 12U) | ((uint32_t) a
[1] << 8U) | ((uint32_t) a
[2] << 4U) | (uint32_t) a
[3];
205 /* Don't allow 0 chars */
215 /* C++11 style 32bit unicode */
221 if (length
!= (size_t) -1 && length
< 9)
224 for (i
= 0; i
< 8; i
++) {
225 a
[i
] = unhexchar(p
[1 + i
]);
230 c
= ((uint32_t) a
[0] << 28U) | ((uint32_t) a
[1] << 24U) | ((uint32_t) a
[2] << 20U) | ((uint32_t) a
[3] << 16U) |
231 ((uint32_t) a
[4] << 12U) | ((uint32_t) a
[5] << 8U) | ((uint32_t) a
[6] << 4U) | (uint32_t) a
[7];
233 /* Don't allow 0 chars */
237 /* Don't allow invalid code points */
238 if (!unichar_is_valid(c
))
258 if (length
!= (size_t) -1 && length
< 3)
273 /* don't allow NUL bytes */
274 if (a
== 0 && b
== 0 && c
== 0)
277 /* Don't allow bytes above 255 */
278 m
= ((uint32_t) a
<< 6U) | ((uint32_t) b
<< 3U) | (uint32_t) c
;
295 int cunescape_length_with_prefix(const char *s
, size_t length
, const char *prefix
, UnescapeFlags flags
, char **ret
) {
303 /* Undoes C style string escaping, and optionally prefixes it. */
305 pl
= strlen_ptr(prefix
);
307 r
= new(char, pl
+length
+1);
312 memcpy(r
, prefix
, pl
);
314 for (f
= s
, t
= r
+ pl
; f
< s
+ length
; f
++) {
316 bool eight_bit
= false;
320 remaining
= s
+ length
- f
;
321 assert(remaining
> 0);
324 /* A literal, copy verbatim */
329 if (remaining
== 1) {
330 if (flags
& UNESCAPE_RELAX
) {
331 /* A trailing backslash, copy verbatim */
340 k
= cunescape_one(f
+ 1, remaining
- 1, &u
, &eight_bit
);
342 if (flags
& UNESCAPE_RELAX
) {
343 /* Invalid escape code, let's take it literal then */
354 /* One byte? Set directly as specified */
357 /* Otherwise encode as multi-byte UTF-8 */
358 t
+= utf8_encode_unichar(t
, u
);
367 int cunescape_length(const char *s
, size_t length
, UnescapeFlags flags
, char **ret
) {
368 return cunescape_length_with_prefix(s
, length
, NULL
, flags
, ret
);
371 int cunescape(const char *s
, UnescapeFlags flags
, char **ret
) {
372 return cunescape_length(s
, strlen(s
), flags
, ret
);
375 char *xescape(const char *s
, const char *bad
) {
379 /* Escapes all chars in bad, in addition to \ and all special
380 * chars, in \xFF style escaping. May be reversed with
383 r
= new(char, strlen(s
) * 4 + 1);
387 for (f
= s
, t
= r
; *f
; f
++) {
389 if ((*f
< ' ') || (*f
>= 127) ||
390 (*f
== '\\') || strchr(bad
, *f
)) {
393 *(t
++) = hexchar(*f
>> 4);
394 *(t
++) = hexchar(*f
);
404 char *octescape(const char *s
, size_t len
) {
408 /* Escapes all chars in bad, in addition to \ and " chars,
409 * in \nnn style escaping. */
411 r
= new(char, len
* 4 + 1);
415 for (f
= s
, t
= r
; f
< s
+ len
; f
++) {
417 if (*f
< ' ' || *f
>= 127 || IN_SET(*f
, '\\', '"')) {
419 *(t
++) = '0' + (*f
>> 6);
420 *(t
++) = '0' + ((*f
>> 3) & 8);
421 *(t
++) = '0' + (*f
& 8);
432 static char *strcpy_backslash_escaped(char *t
, const char *s
, const char *bad
, bool escape_tab_nl
) {
436 if (escape_tab_nl
&& IN_SET(*s
, '\n', '\t')) {
438 *(t
++) = *s
== '\n' ? 'n' : 't';
442 if (*s
== '\\' || strchr(bad
, *s
))
451 char *shell_escape(const char *s
, const char *bad
) {
454 r
= new(char, strlen(s
)*2+1);
458 t
= strcpy_backslash_escaped(r
, s
, bad
, false);
464 char* shell_maybe_quote(const char *s
, EscapeStyle style
) {
470 /* Encloses a string in quotes if necessary to make it OK as a shell
471 * string. Note that we treat benign UTF-8 characters as needing
472 * escaping too, but that should be OK. */
477 strchr(SHELL_NEED_QUOTES
, *p
))
483 r
= new(char, (style
== ESCAPE_POSIX
) + 1 + strlen(s
)*2 + 1 + 1);
488 if (style
== ESCAPE_BACKSLASH
)
490 else if (style
== ESCAPE_POSIX
) {
494 assert_not_reached("Bad EscapeStyle");
496 t
= mempcpy(t
, s
, p
- s
);
498 if (style
== ESCAPE_BACKSLASH
)
499 t
= strcpy_backslash_escaped(t
, p
, SHELL_NEED_ESCAPE
, false);
501 t
= strcpy_backslash_escaped(t
, p
, SHELL_NEED_ESCAPE_POSIX
, true);
503 if (style
== ESCAPE_BACKSLASH
)