15263807472

/* SPDX-License-Identifier: LGPL-2.1-or-later */

#include <stdlib.h>
#include <string.h>

#include "alloc-util.h"
#include "escape.h"
#include "hexdecoct.h"
#include "string-util.h"
#include "strv.h"
#include "utf8.h"

int cescape_char(char c, char *buf) {
        char *buf_old = buf;

        /* Needs space for 4 characters in the buffer */

        switch (c) {

                case '\a':
                        *(buf++) = '\\';
                        *(buf++) = 'a';
                        break;
                case '\b':
                        *(buf++) = '\\';
                        *(buf++) = 'b';
                        break;
                case '\f':
                        *(buf++) = '\\';
                        *(buf++) = 'f';
                        break;
                case '\n':
                        *(buf++) = '\\';
                        *(buf++) = 'n';
                        break;
                case '\r':
                        *(buf++) = '\\';
                        *(buf++) = 'r';
                        break;
                case '\t':
                        *(buf++) = '\\';
                        *(buf++) = 't';
                        break;
                case '\v':
                        *(buf++) = '\\';
                        *(buf++) = 'v';
                        break;
                case '\\':
                        *(buf++) = '\\';
                        *(buf++) = '\\';
                        break;
                case '"':
                        *(buf++) = '\\';
                        *(buf++) = '"';
                        break;
                case '\'':
                        *(buf++) = '\\';
                        *(buf++) = '\'';
                        break;

                default:
                        /* For special chars we prefer octal over
                         * hexadecimal encoding, simply because glib's
                         * g_strescape() does the same */
                        if ((c < ' ') || (c >= 127)) {
                                *(buf++) = '\\';
                                *(buf++) = octchar((unsigned char) c >> 6);
                                *(buf++) = octchar((unsigned char) c >> 3);
                                *(buf++) = octchar((unsigned char) c);
                        } else
                                *(buf++) = c;
                        break;
        }

        return buf - buf_old;
}

char* cescape_length(const char *s, size_t n) {
        const char *f;
        char *r, *t;

        /* Does C style string escaping. May be reversed with cunescape(). */

        assert(s || n == 0);

        if (n == SIZE_MAX)
                n = strlen(s);

        if (n > (SIZE_MAX - 1) / 4)
                return NULL;

        r = new(char, n*4 + 1);
        if (!r)
                return NULL;

        for (f = s, t = r; f < s + n; f++)
                t += cescape_char(*f, t);

        *t = 0;

        return r;
}

int cunescape_one(const char *p, size_t length, char32_t *ret, bool *eight_bit, bool accept_nul) {
        int r = 1;

        assert(p);
        assert(ret);

        /* Unescapes C style. Returns the unescaped character in ret.
         * Sets *eight_bit to true if the escaped sequence either fits in
         * one byte in UTF-8 or is a non-unicode literal byte and should
         * instead be copied directly.
         */

        if (length != SIZE_MAX && length < 1)
                return -EINVAL;

        switch (p[0]) {

        case 'a':
                *ret = '\a';
                break;
        case 'b':
                *ret = '\b';
                break;
        case 'f':
                *ret = '\f';
                break;
        case 'n':
                *ret = '\n';
                break;
        case 'r':
                *ret = '\r';
                break;
        case 't':
                *ret = '\t';
                break;
        case 'v':
                *ret = '\v';
                break;
        case '\\':
                *ret = '\\';
                break;
        case '"':
                *ret = '"';
                break;
        case '\'':
                *ret = '\'';
                break;

        case 's':
                /* This is an extension of the XDG syntax files */
                *ret = ' ';
                break;

        case 'x': {
                /* hexadecimal encoding */
                int a, b;

                if (length != SIZE_MAX && length < 3)
                        return -EINVAL;

                a = unhexchar(p[1]);
                if (a < 0)
                        return -EINVAL;

                b = unhexchar(p[2]);
                if (b < 0)
                        return -EINVAL;

                /* Don't allow NUL bytes */
                if (a == 0 && b == 0 && !accept_nul)
                        return -EINVAL;

                *ret = (a << 4U) | b;
                *eight_bit = true;
                r = 3;
                break;
        }

        case 'u': {
                /* C++11 style 16-bit unicode */

                int a[4];
                size_t i;
                uint32_t c;

                if (length != SIZE_MAX && length < 5)
                        return -EINVAL;

                for (i = 0; i < 4; i++) {
                        a[i] = unhexchar(p[1 + i]);
                        if (a[i] < 0)
                                return a[i];
                }

                c = ((uint32_t) a[0] << 12U) | ((uint32_t) a[1] << 8U) | ((uint32_t) a[2] << 4U) | (uint32_t) a[3];

                /* Don't allow 0 chars */
                if (c == 0 && !accept_nul)
                        return -EINVAL;

                *ret = c;
                r = 5;
                break;
        }

        case 'U': {
                /* C++11 style 32-bit unicode */

                int a[8];
                size_t i;
                char32_t c;

                if (length != SIZE_MAX && length < 9)
                        return -EINVAL;

                for (i = 0; i < 8; i++) {
                        a[i] = unhexchar(p[1 + i]);
                        if (a[i] < 0)
                                return a[i];
                }

                c = ((uint32_t) a[0] << 28U) | ((uint32_t) a[1] << 24U) | ((uint32_t) a[2] << 20U) | ((uint32_t) a[3] << 16U) |
                    ((uint32_t) a[4] << 12U) | ((uint32_t) a[5] <<  8U) | ((uint32_t) a[6] <<  4U) |  (uint32_t) a[7];

                /* Don't allow 0 chars */
                if (c == 0 && !accept_nul)
                        return -EINVAL;

                /* Don't allow invalid code points */
                if (!unichar_is_valid(c))
                        return -EINVAL;

                *ret = c;
                r = 9;
                break;
        }

        case '0':
        case '1':
        case '2':
        case '3':
        case '4':
        case '5':
        case '6':
        case '7': {
                /* octal encoding */
                int a, b, c;
                char32_t m;

                if (length != SIZE_MAX && length < 3)
                        return -EINVAL;

                a = unoctchar(p[0]);
                if (a < 0)
                        return -EINVAL;

                b = unoctchar(p[1]);
                if (b < 0)
                        return -EINVAL;

                c = unoctchar(p[2]);
                if (c < 0)
                        return -EINVAL;

                /* don't allow NUL bytes */
                if (a == 0 && b == 0 && c == 0 && !accept_nul)
                        return -EINVAL;

                /* Don't allow bytes above 255 */
                m = ((uint32_t) a << 6U) | ((uint32_t) b << 3U) | (uint32_t) c;
                if (m > 255)
                        return -EINVAL;

                *ret = m;
                *eight_bit = true;
                r = 3;
                break;
        }

        default:
                return -EINVAL;
        }

        return r;
}

ssize_t cunescape_length_with_prefix(const char *s, size_t length, const char *prefix, UnescapeFlags flags, char **ret) {
        _cleanup_free_ char *ans = NULL;
        char *t;
        const char *f;
        size_t pl;
        int r;

        assert(s);
        assert(ret);

        /* Undoes C style string escaping, and optionally prefixes it. */

        if (length == SIZE_MAX)
                length = strlen(s);

        pl = strlen_ptr(prefix);

        ans = new(char, pl+length+1);
        if (!ans)
                return -ENOMEM;

        if (prefix)
                memcpy(ans, prefix, pl);

        for (f = s, t = ans + pl; f < s + length; f++) {
                size_t remaining;
                bool eight_bit = false;
                char32_t u;

                remaining = s + length - f;
                assert(remaining > 0);

                if (*f != '\\') {
                        /* A literal, copy verbatim */
                        *(t++) = *f;
                        continue;
                }

                if (remaining == 1) {
                        if (flags & UNESCAPE_RELAX) {
                                /* A trailing backslash, copy verbatim */
                                *(t++) = *f;
                                continue;
                        }

                        return -EINVAL;
                }

                r = cunescape_one(f + 1, remaining - 1, &u, &eight_bit, flags & UNESCAPE_ACCEPT_NUL);
                if (r < 0) {
                        if (flags & UNESCAPE_RELAX) {
                                /* Invalid escape code, let's take it literal then */
                                *(t++) = '\\';
                                continue;
                        }

                        return r;
                }

                f += r;
                if (eight_bit)
                        /* One byte? Set directly as specified */
                        *(t++) = u;
                else
                        /* Otherwise encode as multi-byte UTF-8 */
                        t += utf8_encode_unichar(t, u);
        }

        *t = 0;

        assert(t >= ans); /* Let static analyzers know that the answer is non-negative. */
        *ret = TAKE_PTR(ans);
        return t - *ret;
}

char* xescape_full(const char *s, const char *bad, size_t console_width, XEscapeFlags flags) {
        char *ans, *t, *prev, *prev2;
        const char *f;

        assert(s);

        /* Escapes all chars in bad, in addition to \ and all special chars, in \xFF style escaping. May be
         * reversed with cunescape(). If XESCAPE_8_BIT is specified, characters >= 127 are let through
         * unchanged. This corresponds to non-ASCII printable characters in pre-unicode encodings.
         *
         * If console_width is reached, or XESCAPE_FORCE_ELLIPSIS is set, output is truncated and "..." is
         * appended. */

        if (console_width == 0)
                return strdup("");

        ans = new(char, MIN(strlen(s), console_width) * 4 + 1);
        if (!ans)
                return NULL;

        memset(ans, '_', MIN(strlen(s), console_width) * 4);
        ans[MIN(strlen(s), console_width) * 4] = 0;

        bool force_ellipsis = FLAGS_SET(flags, XESCAPE_FORCE_ELLIPSIS);

        for (f = s, t = prev = prev2 = ans; ; f++) {
                char *tmp_t = t;

                if (!*f) {
                        if (force_ellipsis)
                                break;

                        *t = 0;
                        return ans;
                }

                if ((unsigned char) *f < ' ' ||
                    (!FLAGS_SET(flags, XESCAPE_8_BIT) && (unsigned char) *f >= 127) ||
                    *f == '\\' || (bad && strchr(bad, *f))) {
                        if ((size_t) (t - ans) + 4 + 3 * force_ellipsis > console_width)
                                break;

                        *(t++) = '\\';
                        *(t++) = 'x';
                        *(t++) = hexchar(*f >> 4);
                        *(t++) = hexchar(*f);
                } else {
                        if ((size_t) (t - ans) + 1 + 3 * force_ellipsis > console_width)
                                break;

                        *(t++) = *f;
                }

                /* We might need to go back two cycles to fit three dots, so remember two positions */
                prev2 = prev;
                prev = tmp_t;
        }

        /* We can just write where we want, since chars are one-byte */
        size_t c = MIN(console_width, 3u); /* If the console is too narrow, write fewer dots */
        size_t off;
        if (console_width - c >= (size_t) (t - ans))
                off = (size_t) (t - ans);
        else if (console_width - c >= (size_t) (prev - ans))
                off = (size_t) (prev - ans);
        else if (console_width - c >= (size_t) (prev2 - ans))
                off = (size_t) (prev2 - ans);
        else
                off = console_width - c;
        assert(off <= (size_t) (t - ans));

        memcpy(ans + off, "...", c);
        ans[off + c] = '\0';
        return ans;
}

char* escape_non_printable_full(const char *str, size_t console_width, XEscapeFlags flags) {
        if (FLAGS_SET(flags, XESCAPE_8_BIT))
                return xescape_full(str, /* bad= */ NULL, console_width, flags);
        else
                return utf8_escape_non_printable_full(str,
                                                      console_width,
                                                      FLAGS_SET(flags, XESCAPE_FORCE_ELLIPSIS));
}

char* octescape(const char *s, size_t len) {
        char *buf, *t;

        /* Escapes \ and " chars, in \nnn style escaping. */

        assert(s || len == 0);

        if (len == SIZE_MAX)
                len = strlen(s);

        if (len > (SIZE_MAX - 1) / 4)
                return NULL;

        t = buf = new(char, len * 4 + 1);
        if (!buf)
                return NULL;

        for (size_t i = 0; i < len; i++) {
                uint8_t u = (uint8_t) s[i];

                if (u < ' ' || u >= 127 || IN_SET(u, '\\', '"')) {
                        *(t++) = '\\';
                        *(t++) = '0' + (u >> 6);
                        *(t++) = '0' + ((u >> 3) & 7);
                        *(t++) = '0' + (u & 7);
                } else
                        *(t++) = u;
        }

        *t = 0;
        return buf;
}

char* decescape(const char *s, size_t len, const char *bad) {
        char *buf, *t;

        /* Escapes all chars in bad, in addition to \ and " chars, in \nnn decimal style escaping. */

        assert(s || len == 0);

        if (len == SIZE_MAX)
                len = strlen(s);

        if (len > (SIZE_MAX - 1) / 4)
                return NULL;

        t = buf = new(char, len * 4 + 1);
        if (!buf)
                return NULL;

        for (size_t i = 0; i < len; i++) {
                uint8_t u = (uint8_t) s[i];

                if (u < ' ' || u >= 127 || IN_SET(u, '\\', '"') || strchr(bad, u)) {
                        *(t++) = '\\';
                        *(t++) = '0' + (u / 100);
                        *(t++) = '0' + ((u / 10) % 10);
                        *(t++) = '0' + (u % 10);
                } else
                        *(t++) = u;
        }

        *t = 0;
        return buf;
}

static char* strcpy_backslash_escaped(char *t, const char *s, const char *bad) {
        assert(bad);
        assert(t);
        assert(s);

        while (*s) {
                int l = utf8_encoded_valid_unichar(s, SIZE_MAX);

                if (char_is_cc(*s) || l < 0)
                        t += cescape_char(*(s++), t);
                else if (l == 1) {
                        if (*s == '\\' || strchr(bad, *s))
                                *(t++) = '\\';
                        *(t++) = *(s++);
                } else {
                        t = mempcpy(t, s, l);
                        s += l;
                }
        }

        return t;
}

char* shell_escape(const char *s, const char *bad) {
        char *buf, *t;

        buf = new(char, strlen(s)*4+1);
        if (!buf)
                return NULL;

        t = strcpy_backslash_escaped(buf, s, bad);
        *t = 0;

        return buf;
}

char* shell_maybe_quote(const char *s, ShellEscapeFlags flags) {
        const char *p;
        char *buf, *t;

        assert(s);

        /* Encloses a string in quotes if necessary to make it OK as a shell string. */

        if (FLAGS_SET(flags, SHELL_ESCAPE_EMPTY) && isempty(s))
                return strdup("\"\""); /* We don't use $'' here in the POSIX mode. "" is fine too. */

        for (p = s; *p; ) {
                int l = utf8_encoded_valid_unichar(p, SIZE_MAX);

                if (char_is_cc(*p) || l < 0 ||
                    strchr(WHITESPACE SHELL_NEED_QUOTES, *p))
                        break;

                p += l;
        }

        if (!*p)
                return strdup(s);

        buf = new(char, FLAGS_SET(flags, SHELL_ESCAPE_POSIX) + 1 + strlen(s)*4 + 1 + 1);
        if (!buf)
                return NULL;

        t = buf;
        if (FLAGS_SET(flags, SHELL_ESCAPE_POSIX)) {
                *(t++) = '$';
                *(t++) = '\'';
        } else
                *(t++) = '"';

        t = mempcpy(t, s, p - s);

        t = strcpy_backslash_escaped(t, p,
                                     FLAGS_SET(flags, SHELL_ESCAPE_POSIX) ? SHELL_NEED_ESCAPE_POSIX : SHELL_NEED_ESCAPE);

        if (FLAGS_SET(flags, SHELL_ESCAPE_POSIX))
                *(t++) = '\'';
        else
                *(t++) = '"';
        *t = 0;

        return str_realloc(buf);
}

char* quote_command_line(char **argv, ShellEscapeFlags flags) {
        _cleanup_free_ char *result = NULL;

        assert(argv);

        STRV_FOREACH(a, argv) {
                _cleanup_free_ char *t = NULL;

                t = shell_maybe_quote(*a, flags);
                if (!t)
                        return NULL;

                if (!strextend_with_separator(&result, " ", t))
                        return NULL;
        }

        return str_realloc(TAKE_PTR(result));
}

1	/* SPDX-License-Identifier: LGPL-2.1-or-later */
2
3	#include <stdlib.h>
4	#include <string.h>
5
6	#include "alloc-util.h"
7	#include "escape.h"
8	#include "hexdecoct.h"
9	#include "string-util.h"
10	#include "strv.h"
11	#include "utf8.h"
12
13	int cescape_char(char c, char *buf) {	931,051✔
14	char *buf_old = buf;	931,051✔
15
16	/* Needs space for 4 characters in the buffer */
17
18	switch (c) {	931,051✔
19
20	case '\a':	7✔
21	*(buf++) = '\\';	7✔
22	*(buf++) = 'a';	7✔
23	break;	7✔
24	case '\b':	9✔
25	*(buf++) = '\\';	9✔
26	*(buf++) = 'b';	9✔
27	break;	9✔
28	case '\f':	7✔
29	*(buf++) = '\\';	7✔
30	*(buf++) = 'f';	7✔
31	break;	7✔
32	case '\n':	261✔
33	*(buf++) = '\\';	261✔
34	*(buf++) = 'n';	261✔
35	break;	261✔
36	case '\r':	4✔
37	*(buf++) = '\\';	4✔
38	*(buf++) = 'r';	4✔
39	break;	4✔
40	case '\t':	45✔
41	*(buf++) = '\\';	45✔
42	*(buf++) = 't';	45✔
43	break;	45✔
44	case '\v':	2✔
45	*(buf++) = '\\';	2✔
46	*(buf++) = 'v';	2✔
47	break;	2✔
48	case '\\':	611✔
49	*(buf++) = '\\';	611✔
50	*(buf++) = '\\';	611✔
51	break;	611✔
52	case '"':	2,322✔
53	*(buf++) = '\\';	2,322✔
54	*(buf++) = '"';	2,322✔
55	break;	2,322✔
56	case '\'':	19✔
57	*(buf++) = '\\';	19✔
58	*(buf++) = '\'';	19✔
59	break;	19✔
60
61	default:	927,764✔
62	/* For special chars we prefer octal over
63	* hexadecimal encoding, simply because glib's
64	* g_strescape() does the same */
65	if ((c < ' ') \|\| (c >= 127)) {	927,764✔
66	*(buf++) = '\\';	13,823✔
67	*(buf++) = octchar((unsigned char) c >> 6);	13,823✔
68	*(buf++) = octchar((unsigned char) c >> 3);	13,823✔
69	*(buf++) = octchar((unsigned char) c);	13,823✔
70	} else
71	*(buf++) = c;	913,941✔
72	break;
73	}
74
75	return buf - buf_old;	931,051✔
76	}
77
78	char* cescape_length(const char *s, size_t n) {	17,677✔
79	const char *f;	17,677✔
80	char r, t;	17,677✔
81
82	/* Does C style string escaping. May be reversed with cunescape(). */
83
84	assert(s \|\| n == 0);	17,677✔
85
86	if (n == SIZE_MAX)	17,677✔
87	n = strlen(s);	8,067✔
88
89	if (n > (SIZE_MAX - 1) / 4)	17,677✔
90	return NULL;
91
92	r = new(char, n*4 + 1);	17,677✔
93	if (!r)	17,677✔
94	return NULL;
95
96	for (f = s, t = r; f < s + n; f++)	485,803✔
97	t += cescape_char(*f, t);	468,126✔
98
99	*t = 0;	17,677✔
100
101	return r;	17,677✔
102	}
103
104	int cunescape_one(const char p, size_t length, char32_t ret, bool *eight_bit, bool accept_nul) {	6,683✔
105	int r = 1;	6,683✔
106
107	assert(p);	6,683✔
108	assert(ret);	6,683✔
109
110	/* Unescapes C style. Returns the unescaped character in ret.
111	* Sets *eight_bit to true if the escaped sequence either fits in
112	* one byte in UTF-8 or is a non-unicode literal byte and should
113	* instead be copied directly.
114	*/
115
116	if (length != SIZE_MAX && length < 1)	6,683✔
117	return -EINVAL;
118
119	switch (p[0]) {	6,683✔
120
121	case 'a':	4✔
122	*ret = '\a';	4✔
123	break;	4✔
124	case 'b':	8✔
125	*ret = '\b';	8✔
126	break;	8✔
127	case 'f':	3✔
128	*ret = '\f';	3✔
129	break;	3✔
130	case 'n':	22✔
131	*ret = '\n';	22✔
132	break;	22✔
133	case 'r':	4✔
134	*ret = '\r';	4✔
135	break;	4✔
136	case 't':	6✔
137	*ret = '\t';	6✔
138	break;	6✔
139	case 'v':	3✔
140	*ret = '\v';	3✔
141	break;	3✔
142	case '\\':	15✔
143	*ret = '\\';	15✔
144	break;	15✔
145	case '"':	5✔
146	*ret = '"';	5✔
147	break;	5✔
148	case '\'':	1✔
149	*ret = '\'';	1✔
150	break;	1✔
151
152	case 's':	3✔
153	/* This is an extension of the XDG syntax files */
154	*ret = ' ';	3✔
155	break;	3✔
156
157	case 'x': {	6,527✔
158	/* hexadecimal encoding */
159	int a, b;	6,527✔
160
161	if (length != SIZE_MAX && length < 3)	6,527✔
162	return -EINVAL;
163
164	a = unhexchar(p[1]);	6,523✔
165	if (a < 0)	6,523✔
166	return -EINVAL;
167
168	b = unhexchar(p[2]);	6,521✔
169	if (b < 0)	6,521✔
170	return -EINVAL;
171
172	/* Don't allow NUL bytes */
173	if (a == 0 && b == 0 && !accept_nul)	6,521✔
174	return -EINVAL;
175
176	*ret = (a << 4U) \| b;	6,520✔
177	*eight_bit = true;	6,520✔
178	r = 3;	6,520✔
179	break;	6,520✔
180	}
181
182	case 'u': {	19✔
183	/* C++11 style 16-bit unicode */
184
185	int a[4];	19✔
186	size_t i;	19✔
187	uint32_t c;	19✔
188
189	if (length != SIZE_MAX && length < 5)	19✔
190	return -EINVAL;	1✔
191
192	for (i = 0; i < 4; i++) {	95✔
193	a[i] = unhexchar(p[1 + i]);	76✔
194	if (a[i] < 0)	76✔
195	return a[i];
196	}
197
198	c = ((uint32_t) a[0] << 12U) \| ((uint32_t) a[1] << 8U) \| ((uint32_t) a[2] << 4U) \| (uint32_t) a[3];	19✔
199
200	/* Don't allow 0 chars */
201	if (c == 0 && !accept_nul)	19✔
202	return -EINVAL;
203
204	*ret = c;	18✔
205	r = 5;	18✔
206	break;	18✔
207	}
208
209	case 'U': {	10✔
210	/* C++11 style 32-bit unicode */
211
212	int a[8];	10✔
213	size_t i;	10✔
214	char32_t c;	10✔
215
216	if (length != SIZE_MAX && length < 9)	10✔
UNCOV 217	return -EINVAL;	×
218
219	for (i = 0; i < 8; i++) {	90✔
220	a[i] = unhexchar(p[1 + i]);	80✔
221	if (a[i] < 0)	80✔
222	return a[i];
223	}
224
225	c = ((uint32_t) a[0] << 28U) \| ((uint32_t) a[1] << 24U) \| ((uint32_t) a[2] << 20U) \| ((uint32_t) a[3] << 16U) \|	10✔
226	((uint32_t) a[4] << 12U) \| ((uint32_t) a[5] << 8U) \| ((uint32_t) a[6] << 4U) \| (uint32_t) a[7];	10✔
227
228	/* Don't allow 0 chars */
229	if (c == 0 && !accept_nul)	10✔
230	return -EINVAL;
231
232	/* Don't allow invalid code points */
233	if (!unichar_is_valid(c))	10✔
234	return -EINVAL;
235
236	*ret = c;	10✔
237	r = 9;	10✔
238	break;	10✔
239	}
240
241	case '0':	20✔
242	case '1':
243	case '2':
244	case '3':
245	case '4':
246	case '5':
247	case '6':
248	case '7': {
249	/* octal encoding */
250	int a, b, c;	20✔
251	char32_t m;	20✔
252
253	if (length != SIZE_MAX && length < 3)	20✔
254	return -EINVAL;
255
256	a = unoctchar(p[0]);	16✔
257	if (a < 0)	16✔
258	return -EINVAL;
259
260	b = unoctchar(p[1]);	16✔
261	if (b < 0)	16✔
262	return -EINVAL;
263
264	c = unoctchar(p[2]);	16✔
265	if (c < 0)	16✔
266	return -EINVAL;
267
268	/* don't allow NUL bytes */
269	if (a == 0 && b == 0 && c == 0 && !accept_nul)	16✔
270	return -EINVAL;
271
272	/* Don't allow bytes above 255 */
273	m = ((uint32_t) a << 6U) \| ((uint32_t) b << 3U) \| (uint32_t) c;	14✔
274	if (m > 255)	14✔
275	return -EINVAL;
276
277	*ret = m;	14✔
278	*eight_bit = true;	14✔
279	r = 3;	14✔
280	break;	14✔
281	}
282
283	default:
284	return -EINVAL;
285	}
286
287	return r;
288	}
289
290	ssize_t cunescape_length_with_prefix(const char s, size_t length, const char prefix, UnescapeFlags flags, char **ret) {	102,277✔
291	_cleanup_free_ char *ans = NULL;	102,277✔
292	char *t;	102,277✔
293	const char *f;	102,277✔
294	size_t pl;	102,277✔
295	int r;	102,277✔
296
297	assert(s);	102,277✔
298	assert(ret);	102,277✔
299
300	/* Undoes C style string escaping, and optionally prefixes it. */
301
302	if (length == SIZE_MAX)	102,277✔
303	length = strlen(s);	97,944✔
304
305	pl = strlen_ptr(prefix);	102,277✔
306
307	ans = new(char, pl+length+1);	102,277✔
308	if (!ans)	102,277✔
309	return -ENOMEM;
310
311	if (prefix)	102,277✔
312	memcpy(ans, prefix, pl);	4,297✔
313
314	for (f = s, t = ans + pl; f < s + length; f++) {	2,593,380✔
315	size_t remaining;	2,491,112✔
316	bool eight_bit = false;	2,491,112✔
317	char32_t u;	2,491,112✔
318
319	remaining = s + length - f;	2,491,112✔
320	assert(remaining > 0);	2,491,112✔
321
322	if (*f != '\\') {	2,491,112✔
323	/* A literal, copy verbatim */
324	(t++) = f;	2,484,504✔
325	continue;	2,484,511✔
326	}
327
328	if (remaining == 1) {	6,608✔
329	if (flags & UNESCAPE_RELAX) {	3✔
330	/* A trailing backslash, copy verbatim */
331	(t++) = f;	1✔
332	continue;	1✔
333	}
334
335	return -EINVAL;	9✔
336	}
337
338	r = cunescape_one(f + 1, remaining - 1, &u, &eight_bit, flags & UNESCAPE_ACCEPT_NUL);	6,605✔
339	if (r < 0) {	6,605✔
340	if (flags & UNESCAPE_RELAX) {	13✔
341	/* Invalid escape code, let's take it literal then */
342	*(t++) = '\\';	6✔
343	continue;	6✔
344	}
345
346	return r;	7✔
347	}
348
349	f += r;	6,592✔
350	if (eight_bit)	6,592✔
351	/* One byte? Set directly as specified */
352	*(t++) = u;	6,521✔
353	else
354	/* Otherwise encode as multi-byte UTF-8 */
355	t += utf8_encode_unichar(t, u);	71✔
356	}
357
358	*t = 0;	102,268✔
359
360	assert(t >= ans); /* Let static analyzers know that the answer is non-negative. */	102,268✔
361	*ret = TAKE_PTR(ans);	102,268✔
362	return t - *ret;	102,268✔
363	}
364
365	char* xescape_full(const char s, const char bad, size_t console_width, XEscapeFlags flags) {	61,896✔
366	char ans, t, prev, prev2;	61,896✔
367	const char *f;	61,896✔
368
369	assert(s);	61,896✔
370
371	/* Escapes all chars in bad, in addition to \ and all special chars, in \xFF style escaping. May be
372	* reversed with cunescape(). If XESCAPE_8_BIT is specified, characters >= 127 are let through
373	* unchanged. This corresponds to non-ASCII printable characters in pre-unicode encodings.
374	*
375	* If console_width is reached, or XESCAPE_FORCE_ELLIPSIS is set, output is truncated and "..." is
376	* appended. */
377
378	if (console_width == 0)	61,896✔
379	return strdup("");	4✔
380
381	ans = new(char, MIN(strlen(s), console_width) * 4 + 1);	61,892✔
382	if (!ans)	61,892✔
383	return NULL;
384
385	memset(ans, '_', MIN(strlen(s), console_width) * 4);	61,892✔
386	ans[MIN(strlen(s), console_width) * 4] = 0;	61,892✔
387
388	bool force_ellipsis = FLAGS_SET(flags, XESCAPE_FORCE_ELLIPSIS);	61,892✔
389
390	for (f = s, t = prev = prev2 = ans; ; f++) {	61,892✔
391	char *tmp_t = t;	5,045,576✔
392
393	if (!*f) {	5,045,576✔
394	if (force_ellipsis)	61,688✔
395	break;
396
397	*t = 0;	61,675✔
398	return ans;	61,675✔
399	}
400
401	if ((unsigned char) *f < ' ' \|\|	4,983,888✔
402	(!FLAGS_SET(flags, XESCAPE_8_BIT) && (unsigned char) *f >= 127) \|\|	4,982,761✔
403	f == '\\' \|\| (bad && strchr(bad, f))) {	4,982,428✔
404	if ((size_t) (t - ans) + 4 + 3 * force_ellipsis > console_width)	3,894✔
405	break;
406
407	*(t++) = '\\';	3,710✔
408	*(t++) = 'x';	3,710✔
409	(t++) = hexchar(f >> 4);	3,710✔
410	(t++) = hexchar(f);	3,710✔
411	} else {
412	if ((size_t) (t - ans) + 1 + 3 * force_ellipsis > console_width)	4,979,994✔
413	break;
414
415	(t++) = f;	4,979,974✔
416	}
417
418	/* We might need to go back two cycles to fit three dots, so remember two positions */
419	prev2 = prev;	4,983,684✔
420	prev = tmp_t;	4,983,684✔
421	}
422
423	/* We can just write where we want, since chars are one-byte */
424	size_t c = MIN(console_width, 3u); /* If the console is too narrow, write fewer dots */	217✔
425	size_t off;	217✔
426	if (console_width - c >= (size_t) (t - ans))	217✔
427	off = (size_t) (t - ans);
428	else if (console_width - c >= (size_t) (prev - ans))	76✔
429	off = (size_t) (prev - ans);
430	else if (console_width - c >= (size_t) (prev2 - ans))	10✔
431	off = (size_t) (prev2 - ans);
432	else
433	off = console_width - c;	1✔
434	assert(off <= (size_t) (t - ans));	217✔
435
436	memcpy(ans + off, "...", c);	217✔
437	ans[off + c] = '\0';	217✔
438	return ans;	217✔
439	}
440
441	char* escape_non_printable_full(const char *str, size_t console_width, XEscapeFlags flags) {	413✔
442	if (FLAGS_SET(flags, XESCAPE_8_BIT))	413✔
UNCOV 443	return xescape_full(str, /* bad= */ NULL, console_width, flags);	×
444	else
445	return utf8_escape_non_printable_full(str,	413✔
446	console_width,
447	FLAGS_SET(flags, XESCAPE_FORCE_ELLIPSIS));	413✔
448	}
449
450	char* octescape(const char *s, size_t len) {	80✔
451	char buf, t;	80✔
452
453	/* Escapes \ and " chars, in \nnn style escaping. */
454
455	assert(s \|\| len == 0);	80✔
456
457	if (len == SIZE_MAX)	80✔
458	len = strlen(s);	27✔
459
460	if (len > (SIZE_MAX - 1) / 4)	80✔
461	return NULL;
462
463	t = buf = new(char, len * 4 + 1);	80✔
464	if (!buf)	80✔
465	return NULL;
466
467	for (size_t i = 0; i < len; i++) {	908✔
468	uint8_t u = (uint8_t) s[i];	828✔
469
470	if (u < ' ' \|\| u >= 127 \|\| IN_SET(u, '\\', '"')) {	828✔
471	*(t++) = '\\';	5✔
472	*(t++) = '0' + (u >> 6);	5✔
473	*(t++) = '0' + ((u >> 3) & 7);	5✔
474	*(t++) = '0' + (u & 7);	5✔
475	} else
476	*(t++) = u;	823✔
477	}
478
479	*t = 0;	80✔
480	return buf;	80✔
481	}
482
483	char* decescape(const char s, size_t len, const char bad) {	8✔
484	char buf, t;	8✔
485
486	/* Escapes all chars in bad, in addition to \ and " chars, in \nnn decimal style escaping. */
487
488	assert(s \|\| len == 0);	8✔
489
490	if (len == SIZE_MAX)	8✔
491	len = strlen(s);	6✔
492
493	if (len > (SIZE_MAX - 1) / 4)	8✔
494	return NULL;
495
496	t = buf = new(char, len * 4 + 1);	8✔
497	if (!buf)	8✔
498	return NULL;
499
500	for (size_t i = 0; i < len; i++) {	47✔
501	uint8_t u = (uint8_t) s[i];	39✔
502
503	if (u < ' ' \|\| u >= 127 \|\| IN_SET(u, '\\', '"') \|\| strchr(bad, u)) {	39✔
504	*(t++) = '\\';	20✔
505	*(t++) = '0' + (u / 100);	20✔
506	*(t++) = '0' + ((u / 10) % 10);	20✔
507	*(t++) = '0' + (u % 10);	20✔
508	} else
509	*(t++) = u;	19✔
510	}
511
512	*t = 0;	8✔
513	return buf;	8✔
514	}
515
516	static char* strcpy_backslash_escaped(char t, const char s, const char *bad) {	4,860✔
517	assert(bad);	4,860✔
518	assert(t);	4,860✔
519	assert(s);	4,860✔
520
521	while (*s) {	195,659✔
522	int l = utf8_encoded_valid_unichar(s, SIZE_MAX);	190,799✔
523
524	if (char_is_cc(*s) \|\| l < 0)	190,799✔
525	t += cescape_char(*(s++), t);	258✔
526	else if (l == 1) {	190,541✔
527	if (s == '\\' \|\| strchr(bad, s))	190,537✔
528	*(t++) = '\\';	2,202✔
529	(t++) = (s++);	190,537✔
530	} else {
531	t = mempcpy(t, s, l);	4✔
532	s += l;	4✔
533	}
534	}
535
536	return t;	4,860✔
537	}
538
539	char* shell_escape(const char s, const char bad) {	1,433✔
540	char buf, t;	1,433✔
541
542	buf = new(char, strlen(s)*4+1);	1,433✔
543	if (!buf)	1,433✔
544	return NULL;
545
546	t = strcpy_backslash_escaped(buf, s, bad);	1,433✔
547	*t = 0;	1,433✔
548
549	return buf;	1,433✔
550	}
551
552	char* shell_maybe_quote(const char *s, ShellEscapeFlags flags) {	69,628✔
553	const char *p;	69,628✔
554	char buf, t;	69,628✔
555
556	assert(s);	69,628✔
557
558	/* Encloses a string in quotes if necessary to make it OK as a shell string. */
559
560	if (FLAGS_SET(flags, SHELL_ESCAPE_EMPTY) && isempty(s))	69,628✔
561	return strdup("\"\""); /* We don't use $'' here in the POSIX mode. "" is fine too. */	312✔
562
563	for (p = s; *p; ) {	1,503,471✔
564	int l = utf8_encoded_valid_unichar(p, SIZE_MAX);	1,437,582✔
565
566	if (char_is_cc(*p) \|\| l < 0 \|\|	1,437,582✔
567	strchr(WHITESPACE SHELL_NEED_QUOTES, *p))	1,437,555✔
568	break;
569
570	p += l;	1,434,155✔
571	}
572
573	if (!*p)	69,316✔
574	return strdup(s);	65,889✔
575
576	buf = new(char, FLAGS_SET(flags, SHELL_ESCAPE_POSIX) + 1 + strlen(s)*4 + 1 + 1);	6,786✔
577	if (!buf)	3,427✔
578	return NULL;
579
580	t = buf;	3,427✔
581	if (FLAGS_SET(flags, SHELL_ESCAPE_POSIX)) {	3,427✔
582	*(t++) = '$';	68✔
583	*(t++) = '\'';	68✔
584	} else
585	*(t++) = '"';	3,359✔
586
587	t = mempcpy(t, s, p - s);	3,427✔
588
589	t = strcpy_backslash_escaped(t, p,	6,786✔
590	FLAGS_SET(flags, SHELL_ESCAPE_POSIX) ? SHELL_NEED_ESCAPE_POSIX : SHELL_NEED_ESCAPE);
591
592	if (FLAGS_SET(flags, SHELL_ESCAPE_POSIX))	3,427✔
593	*(t++) = '\'';	68✔
594	else
595	*(t++) = '"';	3,359✔
596	*t = 0;	3,427✔
597
598	return str_realloc(buf);	3,427✔
599	}
600
601	char* quote_command_line(char **argv, ShellEscapeFlags flags) {	21,720✔
602	_cleanup_free_ char *result = NULL;	21,720✔
603
604	assert(argv);	21,720✔
605
606	STRV_FOREACH(a, argv) {	90,602✔
607	_cleanup_free_ char *t = NULL;	68,882✔
608
609	t = shell_maybe_quote(*a, flags);	68,882✔
610	if (!t)	68,882✔
611	return NULL;
612
613	if (!strextend_with_separator(&result, " ", t))	68,882✔
614	return NULL;
615	}
616
617	return str_realloc(TAKE_PTR(result));	21,720✔
618	}

systemd / systemd / 15263807472

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous