15263807472

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

#include <fcntl.h>
#include <langinfo.h>
#include <stdlib.h>
#include <sys/mman.h>
#include <sys/stat.h>

#include "dirent-util.h"
#include "env-util.h"
#include "fd-util.h"
#include "fileio.h"
#include "locale-util.h"
#include "log.h"
#include "path-util.h"
#include "process-util.h"
#include "set.h"
#include "string-table.h"
#include "string-util.h"
#include "strv.h"
#include "utf8.h"

static char* normalize_locale(const char *name) {
        const char *e;

        /* Locale names are weird: glibc has some magic rules when looking for the charset name on disk: it
         * lowercases everything, and removes most special chars. This means the official .UTF-8 suffix
         * becomes .utf8 when looking things up on disk. When enumerating locales, let's do the reverse
         * operation, and go back to ".UTF-8" which appears to be the more commonly accepted name. We only do
         * that for UTF-8 however, since it's kinda the only charset that matters. */

        e = endswith(name, ".utf8");
        if (e) {
                _cleanup_free_ char *prefix = NULL;

                prefix = strndup(name, e - name);
                if (!prefix)
                        return NULL;

                return strjoin(prefix, ".UTF-8");
        }

        e = strstr(name, ".utf8@");
        if (e) {
                _cleanup_free_ char *prefix = NULL;

                prefix = strndup(name, e - name);
                if (!prefix)
                        return NULL;

                return strjoin(prefix, ".UTF-8@", e + 6);
        }

        return strdup(name);
}

static int add_locales_from_archive(Set *locales) {
        /* Stolen from glibc... */

        struct locarhead {
                uint32_t magic;
                /* Serial number.  */
                uint32_t serial;
                /* Name hash table.  */
                uint32_t namehash_offset;
                uint32_t namehash_used;
                uint32_t namehash_size;
                /* String table.  */
                uint32_t string_offset;
                uint32_t string_used;
                uint32_t string_size;
                /* Table with locale records.  */
                uint32_t locrectab_offset;
                uint32_t locrectab_used;
                uint32_t locrectab_size;
                /* MD5 sum hash table.  */
                uint32_t sumhash_offset;
                uint32_t sumhash_used;
                uint32_t sumhash_size;
        };

        struct namehashent {
                /* Hash value of the name.  */
                uint32_t hashval;
                /* Offset of the name in the string table.  */
                uint32_t name_offset;
                /* Offset of the locale record.  */
                uint32_t locrec_offset;
        };

        int r;

        assert(locales);

        _cleanup_close_ int fd = open("/usr/lib/locale/locale-archive", O_RDONLY|O_NOCTTY|O_CLOEXEC);
        if (fd < 0)
                return errno == ENOENT ? 0 : -errno;

        struct stat st;
        if (fstat(fd, &st) < 0)
                return -errno;

        if (!S_ISREG(st.st_mode))
                return -EBADMSG;

        if (st.st_size < (off_t) sizeof(struct locarhead))
                return -EBADMSG;

        if (file_offset_beyond_memory_size(st.st_size))
                return -EFBIG;

        void *p = mmap(NULL, st.st_size, PROT_READ, MAP_SHARED, fd, 0);
        if (p == MAP_FAILED)
                return -errno;

        const struct namehashent *e;
        const struct locarhead *h = p;
        if (h->magic != 0xde020109 ||
            h->namehash_offset + h->namehash_size > st.st_size ||
            h->string_offset + h->string_size > st.st_size ||
            h->locrectab_offset + h->locrectab_size > st.st_size ||
            h->sumhash_offset + h->sumhash_size > st.st_size) {
                r = -EBADMSG;
                goto finish;
        }

        e = (const struct namehashent*) ((const uint8_t*) p + h->namehash_offset);
        for (size_t i = 0; i < h->namehash_size; i++) {
                char *z;

                if (e[i].locrec_offset == 0)
                        continue;

                if (!utf8_is_valid((char*) p + e[i].name_offset))
                        continue;

                z = normalize_locale((char*) p + e[i].name_offset);
                if (!z) {
                        r = -ENOMEM;
                        goto finish;
                }

                r = set_consume(locales, z);
                if (r < 0)
                        goto finish;
        }

        r = 0;

finish:
        if (p != MAP_FAILED)
                munmap((void*) p, st.st_size);

        return r;
}

static int add_locales_from_libdir(Set *locales) {
        _cleanup_closedir_ DIR *dir = NULL;
        int r;

        assert(locales);

        dir = opendir("/usr/lib/locale");
        if (!dir)
                return errno == ENOENT ? 0 : -errno;

        FOREACH_DIRENT(de, dir, return -errno) {
                char *z;

                if (de->d_type != DT_DIR)
                        continue;

                z = normalize_locale(de->d_name);
                if (!z)
                        return -ENOMEM;

                r = set_consume(locales, z);
                if (r < 0)
                        return r;
        }

        return 0;
}

int get_locales(char ***ret) {
        _cleanup_set_free_ Set *locales = NULL;
        int r;

        locales = set_new(&string_hash_ops_free);
        if (!locales)
                return -ENOMEM;

        r = add_locales_from_archive(locales);
        if (r < 0 && r != -ENOENT)
                return r;

        r = add_locales_from_libdir(locales);
        if (r < 0)
                return r;

        char *locale;
        SET_FOREACH(locale, locales) {
                r = locale_is_installed(locale);
                if (r < 0)
                        return r;
                if (r == 0)
                        free(set_remove(locales, locale));
        }

        _cleanup_strv_free_ char **l = set_to_strv(&locales);
        if (!l)
                return -ENOMEM;

        r = getenv_bool("SYSTEMD_LIST_NON_UTF8_LOCALES");
        if (r <= 0) {
                if (!IN_SET(r, -ENXIO, 0))
                        log_debug_errno(r, "Failed to parse $SYSTEMD_LIST_NON_UTF8_LOCALES as boolean, ignoring: %m");

                /* Filter out non-UTF-8 locales, because it's 2019, by default */
                char **b = l;
                STRV_FOREACH(a, l)
                        if (endswith(*a, "UTF-8") || strstr(*a, ".UTF-8@"))
                                *(b++) = *a;
                        else
                                free(*a);

                *b = NULL;
        }

        strv_sort(l);

        *ret = TAKE_PTR(l);

        return 0;
}

bool locale_is_valid(const char *name) {

        if (isempty(name))
                return false;

        if (strlen(name) >= 128)
                return false;

        if (!utf8_is_valid(name))
                return false;

        if (!filename_is_valid(name))
                return false;

        /* Locales look like: ll_CC.ENC@variant, where ll and CC are alphabetic, ENC is alphanumeric with
         * dashes, and variant seems to be alphabetic.
         * See: https://www.gnu.org/software/gettext/manual/html_node/Locale-Names.html */
        if (!in_charset(name, ALPHANUMERICAL "_.-@"))
                return false;

        return true;
}

int locale_is_installed(const char *name) {
        if (!locale_is_valid(name))
                return false;

        if (STR_IN_SET(name, "C", "POSIX")) /* These ones are always OK */
                return true;

        _cleanup_(freelocalep) locale_t loc = newlocale(LC_ALL_MASK, name, (locale_t) 0);
        if (loc == (locale_t) 0)
                return errno == ENOMEM ? -ENOMEM : false;

        return true;
}

static bool is_locale_utf8_impl(void) {
        const char *set;
        int r;

        /* Note that we default to 'true' here, since today UTF8 is pretty much supported everywhere. */

        r = secure_getenv_bool("SYSTEMD_UTF8");
        if (r >= 0)
                return r;
        if (r != -ENXIO)
                log_debug_errno(r, "Failed to parse $SYSTEMD_UTF8, ignoring: %m");

        /* This function may be called from libsystemd, and setlocale() is not thread safe. Assuming yes. */
        if (!is_main_thread())
                return true;

        if (!setlocale(LC_ALL, ""))
                return true;

        set = nl_langinfo(CODESET);
        if (!set || streq(set, "UTF-8"))
                return true;

        set = setlocale(LC_CTYPE, NULL);
        if (!set)
                return true;

        /* Unless LC_CTYPE is explicitly overridden, return true. Because here CTYPE is effectively unset
         * and everything can do to UTF-8 nowadays. */
        return STR_IN_SET(set, "C", "POSIX") &&
                !getenv("LC_ALL") &&
                !getenv("LC_CTYPE") &&
                !getenv("LANG");
}

bool is_locale_utf8(void) {
        static int cached = -1;

        if (cached < 0)
                cached = is_locale_utf8_impl();

        return cached;
}

void locale_variables_free(char *l[_VARIABLE_LC_MAX]) {
        free_many_charp(l, _VARIABLE_LC_MAX);
}

void locale_variables_simplify(char *l[_VARIABLE_LC_MAX]) {
        assert(l);

        for (LocaleVariable p = 0; p < _VARIABLE_LC_MAX; p++) {
                if (p == VARIABLE_LANG)
                        continue;
                if (isempty(l[p]) || streq_ptr(l[VARIABLE_LANG], l[p]))
                        l[p] = mfree(l[p]);
        }
}

static const char * const locale_variable_table[_VARIABLE_LC_MAX] = {
        [VARIABLE_LANG]              = "LANG",
        [VARIABLE_LANGUAGE]          = "LANGUAGE",
        [VARIABLE_LC_CTYPE]          = "LC_CTYPE",
        [VARIABLE_LC_NUMERIC]        = "LC_NUMERIC",
        [VARIABLE_LC_TIME]           = "LC_TIME",
        [VARIABLE_LC_COLLATE]        = "LC_COLLATE",
        [VARIABLE_LC_MONETARY]       = "LC_MONETARY",
        [VARIABLE_LC_MESSAGES]       = "LC_MESSAGES",
        [VARIABLE_LC_PAPER]          = "LC_PAPER",
        [VARIABLE_LC_NAME]           = "LC_NAME",
        [VARIABLE_LC_ADDRESS]        = "LC_ADDRESS",
        [VARIABLE_LC_TELEPHONE]      = "LC_TELEPHONE",
        [VARIABLE_LC_MEASUREMENT]    = "LC_MEASUREMENT",
        [VARIABLE_LC_IDENTIFICATION] = "LC_IDENTIFICATION"
};

DEFINE_STRING_TABLE_LOOKUP(locale_variable, LocaleVariable);

1	/* SPDX-License-Identifier: LGPL-2.1-or-later */
2
3	#include <fcntl.h>
4	#include <langinfo.h>
5	#include <stdlib.h>
6	#include <sys/mman.h>
7	#include <sys/stat.h>
8
9	#include "dirent-util.h"
10	#include "env-util.h"
11	#include "fd-util.h"
12	#include "fileio.h"
13	#include "locale-util.h"
14	#include "log.h"
15	#include "path-util.h"
16	#include "process-util.h"
17	#include "set.h"
18	#include "string-table.h"
19	#include "string-util.h"
20	#include "strv.h"
21	#include "utf8.h"
22
23	static char* normalize_locale(const char *name) {	21✔
24	const char *e;	21✔
25
26	/* Locale names are weird: glibc has some magic rules when looking for the charset name on disk: it
27	* lowercases everything, and removes most special chars. This means the official .UTF-8 suffix
28	* becomes .utf8 when looking things up on disk. When enumerating locales, let's do the reverse
29	* operation, and go back to ".UTF-8" which appears to be the more commonly accepted name. We only do
30	* that for UTF-8 however, since it's kinda the only charset that matters. */
31
32	e = endswith(name, ".utf8");	21✔
33	if (e) {	21✔
34	_cleanup_free_ char *prefix = NULL;	16✔
35
36	prefix = strndup(name, e - name);	16✔
37	if (!prefix)	16✔
38	return NULL;
39
40	return strjoin(prefix, ".UTF-8");	16✔
41	}
42
43	e = strstr(name, ".utf8@");	5✔
44	if (e) {	5✔
UNCOV 45	_cleanup_free_ char *prefix = NULL;	×
46
UNCOV 47	prefix = strndup(name, e - name);	×
UNCOV 48	if (!prefix)	×
49	return NULL;
50
51	return strjoin(prefix, ".UTF-8@", e + 6);	×
52	}
53
54	return strdup(name);	5✔
55	}
56
57	static int add_locales_from_archive(Set *locales) {	9✔
58	/* Stolen from glibc... */
59
60	struct locarhead {	9✔
61	uint32_t magic;
62	/* Serial number. */
63	uint32_t serial;
64	/* Name hash table. */
65	uint32_t namehash_offset;
66	uint32_t namehash_used;
67	uint32_t namehash_size;
68	/* String table. */
69	uint32_t string_offset;
70	uint32_t string_used;
71	uint32_t string_size;
72	/* Table with locale records. */
73	uint32_t locrectab_offset;
74	uint32_t locrectab_used;
75	uint32_t locrectab_size;
76	/* MD5 sum hash table. */
77	uint32_t sumhash_offset;
78	uint32_t sumhash_used;
79	uint32_t sumhash_size;
80	};
81
82	struct namehashent {	9✔
83	/* Hash value of the name. */
84	uint32_t hashval;
85	/* Offset of the name in the string table. */
86	uint32_t name_offset;
87	/* Offset of the locale record. */
88	uint32_t locrec_offset;
89	};
90
91	int r;	9✔
92
93	assert(locales);	9✔
94
95	_cleanup_close_ int fd = open("/usr/lib/locale/locale-archive", O_RDONLY\|O_NOCTTY\|O_CLOEXEC);	18✔
96	if (fd < 0)	9✔
97	return errno == ENOENT ? 0 : -errno;	3✔
98
99	struct stat st;	6✔
100	if (fstat(fd, &st) < 0)	6✔
UNCOV 101	return -errno;	×
102
103	if (!S_ISREG(st.st_mode))	6✔
104	return -EBADMSG;
105
106	if (st.st_size < (off_t) sizeof(struct locarhead))	6✔
107	return -EBADMSG;
108
109	if (file_offset_beyond_memory_size(st.st_size))	6✔
110	return -EFBIG;
111
112	void *p = mmap(NULL, st.st_size, PROT_READ, MAP_SHARED, fd, 0);	6✔
113	if (p == MAP_FAILED)	6✔
UNCOV 114	return -errno;	×
115
116	const struct namehashent *e;	6✔
117	const struct locarhead *h = p;	6✔
118	if (h->magic != 0xde020109 \|\|	6✔
119	h->namehash_offset + h->namehash_size > st.st_size \|\|	6✔
120	h->string_offset + h->string_size > st.st_size \|\|	6✔
121	h->locrectab_offset + h->locrectab_size > st.st_size \|\|	6✔
122	h->sumhash_offset + h->sumhash_size > st.st_size) {	6✔
UNCOV 123	r = -EBADMSG;	×
UNCOV 124	goto finish;	×
125	}
126
127	e = (const struct namehashent) ((const uint8_t) p + h->namehash_offset);	6✔
128	for (size_t i = 0; i < h->namehash_size; i++) {	5,448✔
129	char *z;	5,442✔
130
131	if (e[i].locrec_offset == 0)	5,442✔
132	continue;	5,435✔
133
134	if (!utf8_is_valid((char*) p + e[i].name_offset))	7✔
UNCOV 135	continue;	×
136
137	z = normalize_locale((char*) p + e[i].name_offset);	7✔
138	if (!z) {	7✔
UNCOV 139	r = -ENOMEM;	×
UNCOV 140	goto finish;	×
141	}
142
143	r = set_consume(locales, z);	7✔
144	if (r < 0)	7✔
145	goto finish;	×
146	}
147
148	r = 0;
149
150	finish:	6✔
151	if (p != MAP_FAILED)	6✔
152	munmap((void*) p, st.st_size);	6✔
153
154	return r;	6✔
155	}
156
157	static int add_locales_from_libdir(Set *locales) {	9✔
158	_cleanup_closedir_ DIR *dir = NULL;	9✔
159	int r;	9✔
160
161	assert(locales);	9✔
162
163	dir = opendir("/usr/lib/locale");	9✔
164	if (!dir)	9✔
UNCOV 165	return errno == ENOENT ? 0 : -errno;	×
166
167	FOREACH_DIRENT(de, dir, return -errno) {	47✔
168	char *z;	20✔
169
170	if (de->d_type != DT_DIR)	20✔
171	continue;	6✔
172
173	z = normalize_locale(de->d_name);	14✔
174	if (!z)	14✔
175	return -ENOMEM;
176
177	r = set_consume(locales, z);	14✔
178	if (r < 0)	14✔
179	return r;
180	}
181
182	return 0;
183	}
184
185	int get_locales(char ***ret) {	9✔
186	_cleanup_set_free_ Set *locales = NULL;	9✔
187	int r;	9✔
188
189	locales = set_new(&string_hash_ops_free);	9✔
190	if (!locales)	9✔
191	return -ENOMEM;
192
193	r = add_locales_from_archive(locales);	9✔
194	if (r < 0 && r != -ENOENT)	9✔
195	return r;
196
197	r = add_locales_from_libdir(locales);	9✔
198	if (r < 0)	9✔
199	return r;
200
201	char *locale;	9✔
202	SET_FOREACH(locale, locales) {	39✔
203	r = locale_is_installed(locale);	21✔
204	if (r < 0)	21✔
UNCOV 205	return r;	×
206	if (r == 0)	21✔
207	free(set_remove(locales, locale));	5✔
208	}
209
210	_cleanup_strv_free_ char **l = set_to_strv(&locales);	18✔
211	if (!l)	9✔
212	return -ENOMEM;
213
214	r = getenv_bool("SYSTEMD_LIST_NON_UTF8_LOCALES");	9✔
215	if (r <= 0) {	9✔
216	if (!IN_SET(r, -ENXIO, 0))	9✔
UNCOV 217	log_debug_errno(r, "Failed to parse $SYSTEMD_LIST_NON_UTF8_LOCALES as boolean, ignoring: %m");	×
218
219	/* Filter out non-UTF-8 locales, because it's 2019, by default */
220	char **b = l;	9✔
221	STRV_FOREACH(a, l)	25✔
222	if (endswith(a, "UTF-8") \|\| strstr(a, ".UTF-8@"))	16✔
223	(b++) = a;	16✔
224	else
UNCOV 225	free(*a);	×
226
227	*b = NULL;	9✔
228	}
229
230	strv_sort(l);	9✔
231
232	*ret = TAKE_PTR(l);	9✔
233
234	return 0;	9✔
235	}
236
237	bool locale_is_valid(const char *name) {	73✔
238
239	if (isempty(name))	73✔
240	return false;
241
242	if (strlen(name) >= 128)	71✔
243	return false;
244
245	if (!utf8_is_valid(name))	71✔
246	return false;
247
248	if (!filename_is_valid(name))	71✔
249	return false;
250
251	/* Locales look like: ll_CC.ENC@variant, where ll and CC are alphabetic, ENC is alphanumeric with
252	* dashes, and variant seems to be alphabetic.
253	* See: https://www.gnu.org/software/gettext/manual/html_node/Locale-Names.html */
254	if (!in_charset(name, ALPHANUMERICAL "_.-@"))	69✔
255	return false;	2✔
256
257	return true;
258	}
259
260	int locale_is_installed(const char *name) {	42✔
261	if (!locale_is_valid(name))	42✔
262	return false;	42✔
263
264	if (STR_IN_SET(name, "C", "POSIX")) /* These ones are always OK */	39✔
265	return true;	5✔
266
267	_cleanup_(freelocalep) locale_t loc = newlocale(LC_ALL_MASK, name, (locale_t) 0);	34✔
268	if (loc == (locale_t) 0)	34✔
269	return errno == ENOMEM ? -ENOMEM : false;	12✔
270
271	return true;
272	}
273
274	static bool is_locale_utf8_impl(void) {	12,409✔
275	const char *set;	12,409✔
276	int r;	12,409✔
277
278	/* Note that we default to 'true' here, since today UTF8 is pretty much supported everywhere. */
279
280	r = secure_getenv_bool("SYSTEMD_UTF8");	12,409✔
281	if (r >= 0)	12,409✔
282	return r;	94✔
283	if (r != -ENXIO)	12,315✔
UNCOV 284	log_debug_errno(r, "Failed to parse $SYSTEMD_UTF8, ignoring: %m");	×
285
286	/* This function may be called from libsystemd, and setlocale() is not thread safe. Assuming yes. */
287	if (!is_main_thread())	12,315✔
288	return true;
289
290	if (!setlocale(LC_ALL, ""))	12,293✔
291	return true;
292
293	set = nl_langinfo(CODESET);	12,293✔
294	if (!set \|\| streq(set, "UTF-8"))	12,293✔
295	return true;
296
297	set = setlocale(LC_CTYPE, NULL);	5,163✔
298	if (!set)	5,163✔
299	return true;
300
301	/* Unless LC_CTYPE is explicitly overridden, return true. Because here CTYPE is effectively unset
302	* and everything can do to UTF-8 nowadays. */
303	return STR_IN_SET(set, "C", "POSIX") &&	10,326✔
304	!getenv("LC_ALL") &&	10,326✔
305	!getenv("LC_CTYPE") &&	15,489✔
306	!getenv("LANG");	5,162✔
307	}
308
309	bool is_locale_utf8(void) {	892,419✔
310	static int cached = -1;	892,419✔
311
312	if (cached < 0)	892,419✔
313	cached = is_locale_utf8_impl();	12,409✔
314
315	return cached;	892,419✔
316	}
317
318	void locale_variables_free(char *l[_VARIABLE_LC_MAX]) {	8✔
319	free_many_charp(l, _VARIABLE_LC_MAX);	8✔
320	}	8✔
321
322	void locale_variables_simplify(char *l[_VARIABLE_LC_MAX]) {	278✔
323	assert(l);	278✔
324
325	for (LocaleVariable p = 0; p < _VARIABLE_LC_MAX; p++) {	4,170✔
326	if (p == VARIABLE_LANG)	3,892✔
327	continue;	278✔
328	if (isempty(l[p]) \|\| streq_ptr(l[VARIABLE_LANG], l[p]))	3,629✔
329	l[p] = mfree(l[p]);	3,601✔
330	}
331	}	278✔
332
333	static const char * const locale_variable_table[_VARIABLE_LC_MAX] = {
334	[VARIABLE_LANG] = "LANG",
335	[VARIABLE_LANGUAGE] = "LANGUAGE",
336	[VARIABLE_LC_CTYPE] = "LC_CTYPE",
337	[VARIABLE_LC_NUMERIC] = "LC_NUMERIC",
338	[VARIABLE_LC_TIME] = "LC_TIME",
339	[VARIABLE_LC_COLLATE] = "LC_COLLATE",
340	[VARIABLE_LC_MONETARY] = "LC_MONETARY",
341	[VARIABLE_LC_MESSAGES] = "LC_MESSAGES",
342	[VARIABLE_LC_PAPER] = "LC_PAPER",
343	[VARIABLE_LC_NAME] = "LC_NAME",
344	[VARIABLE_LC_ADDRESS] = "LC_ADDRESS",
345	[VARIABLE_LC_TELEPHONE] = "LC_TELEPHONE",
346	[VARIABLE_LC_MEASUREMENT] = "LC_MEASUREMENT",
347	[VARIABLE_LC_IDENTIFICATION] = "LC_IDENTIFICATION"
348	};
349
350	DEFINE_STRING_TABLE_LOOKUP(locale_variable, LocaleVariable);	15,724✔

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