src/basic/locale-util.c

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