locale/localeinfo.h

   1 /* Declarations for internal libc locale interfaces
   2    Copyright (C) 1995-2016 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <http://www.gnu.org/licenses/>.  */
  18
  19 #ifndef _LOCALEINFO_H
  20 #define _LOCALEINFO_H 1
  21
  22 #include <stddef.h>
  23 #include <langinfo.h>
  24 #include <limits.h>
  25 #include <locale.h>
  26 #include <time.h>
  27 #include <stdint.h>
  28 #include <sys/types.h>
  29
  30 #include <intl/loadinfo.h>      /* For loaded_l10nfile definition.  */
  31
  32 /* Magic number at the beginning of a locale data file for CATEGORY.  */
  33 #define LIMAGIC(category) \
  34   (category == LC_COLLATE                                               \
  35    ? ((unsigned int) (0x20051014 ^ (category)))                         \
  36    : category == LC_CTYPE                                               \
  37    ? ((unsigned int) (0x20090720 ^ (category)))                         \
  38    : ((unsigned int) (0x20031115 ^ (category))))
  39
  40 /* Two special weight constants for the collation data.  */
  41 #define IGNORE_CHAR     2
  42
  43 /* We use a special value for the usage counter in `__locale_data' to
  44    signal that this data must never be removed anymore.  */
  45 #define MAX_USAGE_COUNT (UINT_MAX - 1)
  46 #define UNDELETABLE     UINT_MAX
  47
  48 /* Structure describing locale data in core for a category.  */
  49 struct __locale_data
  50 {
  51   const char *name;
  52   const char *filedata;         /* Region mapping the file data.  */
  53   off_t filesize;               /* Size of the file (and the region).  */
  54   enum                          /* Flavor of storage used for those.  */
  55   {
  56     ld_malloced,                /* Both are malloc'd.  */
  57     ld_mapped,                  /* name is malloc'd, filedata mmap'd */
  58     ld_archive                  /* Both point into mmap'd archive regions.  */
  59   } alloc;
  60
  61   /* This provides a slot for category-specific code to cache data computed
  62      about this locale.  That code can set a cleanup function to deallocate
  63      the data.  */
  64   struct
  65   {
  66     void (*cleanup) (struct __locale_data *) internal_function;
  67     union
  68     {
  69       void *data;
  70       struct lc_time_data *time;
  71       const struct gconv_fcts *ctype;
  72     };
  73   } private;
  74
  75   unsigned int usage_count;     /* Counter for users.  */
  76
  77   int use_translit;             /* Nonzero if the mb*towv*() and wc*tomb()
  78                                    functions should use transliteration.  */
  79
  80   unsigned int nstrings;        /* Number of strings below.  */
  81   union locale_data_value
  82   {
  83     const uint32_t *wstr;
  84     const char *string;
  85     unsigned int word;          /* Note endian issues vs 64-bit pointers.  */
  86   }
  87   values __flexarr;     /* Items, usually pointers into `filedata'.  */
  88 };
  89
  90 /* This alignment is used for 32-bit integers in locale files, both
  91    those that are explicitly int32_t or uint32_t and those that are
  92    wchar_t, regardless of the (possibly smaller) alignment required
  93    for such integers on a particular host.  */
  94 #define LOCFILE_ALIGN           sizeof (int32_t)
  95 #define LOCFILE_ALIGN_MASK      (LOCFILE_ALIGN - 1)
  96 #define LOCFILE_ALIGN_UP(x)     (((x) + LOCFILE_ALIGN - 1)      \
  97                                  & ~LOCFILE_ALIGN_MASK)
  98 #define LOCFILE_ALIGNED_P(x)    (((x) & LOCFILE_ALIGN_MASK) == 0)
  99
 100 /* We know three kinds of collation sorting rules.  */
 101 enum coll_sort_rule
 102 {
 103   illegal_0__,
 104   sort_forward,
 105   sort_backward,
 106   illegal_3__,
 107   sort_position,
 108   sort_forward_position,
 109   sort_backward_position,
 110   sort_mask
 111 };
 112
 113 /* We can map the types of the entries into a few categories.  */
 114 enum value_type
 115 {
 116   none,
 117   string,
 118   stringarray,
 119   byte,
 120   bytearray,
 121   word,
 122   stringlist,
 123   wordarray,
 124   wstring,
 125   wstringarray,
 126   wstringlist
 127 };
 128
 129
 130 /* Definitions for `era' information from LC_TIME.  */
 131 #define ERA_NAME_FORMAT_MEMBERS 4
 132 #define ERA_M_NAME   0
 133 #define ERA_M_FORMAT 1
 134 #define ERA_W_NAME   2
 135 #define ERA_W_FORMAT 3
 136
 137
 138 /* Structure to access `era' information from LC_TIME.  */
 139 struct era_entry
 140 {
 141   uint32_t direction;           /* Contains '+' or '-'.  */
 142   int32_t offset;
 143   int32_t start_date[3];
 144   int32_t stop_date[3];
 145   const char *era_name;
 146   const char *era_format;
 147   const wchar_t *era_wname;
 148   const wchar_t *era_wformat;
 149   int absolute_direction;
 150   /* absolute direction:
 151      +1 indicates that year number is higher in the future. (like A.D.)
 152      -1 indicates that year number is higher in the past. (like B.C.)  */
 153 };
 154
 155 /* Structure caching computed data about information from LC_TIME.
 156    The `private.time' member of `struct __locale_data' points to this.  */
 157 struct lc_time_data
 158 {
 159   struct era_entry *eras;
 160   size_t num_eras;
 161   int era_initialized;
 162
 163   const char **alt_digits;
 164   const wchar_t **walt_digits;
 165   int alt_digits_initialized;
 166   int walt_digits_initialized;
 167 };
 168
 169
 170 /* LC_CTYPE specific:
 171    Hardwired indices for standard wide character translation mappings.  */
 172 enum
 173 {
 174   __TOW_toupper = 0,
 175   __TOW_tolower = 1
 176 };
 177
 178
 179 /* LC_CTYPE specific:
 180    Access a wide character class with a single character index.
 181    _ISCTYPE (c, desc) = iswctype (btowc (c), desc).
 182    c must be an `unsigned char'.  desc must be a nonzero wctype_t.  */
 183 #define _ISCTYPE(c, desc) \
 184   (((((const uint32_t *) (desc)) - 8)[(c) >> 5] >> ((c) & 0x1f)) & 1)
 185
 186 /* Category name handling variables.  */
 187 #define CATNAMEMF(line) CATNAMEMF1 (line)
 188 #define CATNAMEMF1(line) str##line
 189 extern const union catnamestr_t
 190 {
 191   struct
 192   {
 193 #define DEFINE_CATEGORY(category, category_name, items, a) \
 194     char CATNAMEMF (__LINE__)[sizeof (category_name)];
 195 #include "categories.def"
 196 #undef DEFINE_CATEGORY
 197   };
 198   char str[0];
 199 } _nl_category_names attribute_hidden;
 200 extern const uint8_t _nl_category_name_idxs[__LC_LAST] attribute_hidden;
 201 extern const uint8_t _nl_category_name_sizes[__LC_LAST] attribute_hidden;
 202
 203 /* Name of the standard locales.  */
 204 extern const char _nl_C_name[] attribute_hidden;
 205 extern const char _nl_POSIX_name[] attribute_hidden;
 206
 207 /* The standard codeset.  */
 208 extern const char _nl_C_codeset[] attribute_hidden;
 209
 210 /* This is the internal locale_t object that holds the global locale
 211    controlled by calls to setlocale.  A thread's TSD locale pointer
 212    points to this when `uselocale (LC_GLOBAL_LOCALE)' is in effect.  */
 213 extern struct __locale_struct _nl_global_locale attribute_hidden;
 214
 215 /* This fetches the thread-local locale_t pointer, either one set with
 216    uselocale or &_nl_global_locale.  */
 217 #define _NL_CURRENT_LOCALE      (__libc_tsd_get (__locale_t, LOCALE))
 218 #include <libc-tsd.h>
 219 __libc_tsd_define (extern, __locale_t, LOCALE)
 220
 221
 222 /* For static linking it is desireable to avoid always linking in the code
 223    and data for every category when we can tell at link time that they are
 224    unused.  We can manage this playing some tricks with weak references.
 225    But with thread-local locale settings, it becomes quite ungainly unless
 226    we can use __thread variables.  So only in that case do we attempt this.  */
 227 #ifndef SHARED
 228 # include <tls.h>
 229 # define NL_CURRENT_INDIRECT    1
 230 #endif
 231
 232 #ifdef NL_CURRENT_INDIRECT
 233
 234 /* For each category declare the thread-local variable for the current
 235    locale data.  This has an extra indirection so it points at the
 236    __locales[CATEGORY] element in either _nl_global_locale or the current
 237    locale object set by uselocale, which points at the actual data.  The
 238    reason for having these variables is so that references to particular
 239    categories will link in the lc-CATEGORY.c module to define this symbol,
 240    and we arrange that linking that module is what brings in all the code
 241    associated with this category.  */
 242 #define DEFINE_CATEGORY(category, category_name, items, a) \
 243 extern __thread struct __locale_data *const *_nl_current_##category \
 244   attribute_hidden attribute_tls_model_ie;
 245 #include "categories.def"
 246 #undef  DEFINE_CATEGORY
 247
 248 /* Return a pointer to the current `struct __locale_data' for CATEGORY.  */
 249 #define _NL_CURRENT_DATA(category)      (*_nl_current_##category)
 250
 251 /* Extract the current CATEGORY locale's string for ITEM.  */
 252 #define _NL_CURRENT(category, item) \
 253   ((*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].string)
 254
 255 /* Extract the current CATEGORY locale's string for ITEM.  */
 256 #define _NL_CURRENT_WSTR(category, item) \
 257   ((wchar_t *) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].wstr)
 258
 259 /* Extract the current CATEGORY locale's word for ITEM.  */
 260 #define _NL_CURRENT_WORD(category, item) \
 261   ((uint32_t) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].word)
 262
 263 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  The symbol
 264    _nl_current_CATEGORY_used is set to a value unequal to zero to mark this
 265    category as used.  On S390 the used relocation to load the symbol address
 266    can only handle even addresses.  */
 267 #define _NL_CURRENT_DEFINE(category) \
 268   __thread struct __locale_data *const *_nl_current_##category \
 269     attribute_hidden = &_nl_global_locale.__locales[category]; \
 270   asm (".globl " __SYMBOL_PREFIX "_nl_current_" #category "_used\n" \
 271        _NL_CURRENT_DEFINE_ABS (_nl_current_##category##_used, 2));
 272 #ifdef HAVE_ASM_SET_DIRECTIVE
 273 # define _NL_CURRENT_DEFINE_ABS(sym, val) ".set " #sym ", " #val
 274 #else
 275 # define _NL_CURRENT_DEFINE_ABS(sym, val) #sym " = " #val
 276 #endif
 277
 278 #else
 279
 280 /* All categories are always loaded in the shared library, so there is no
 281    point in having lots of separate symbols for linking.  */
 282
 283 /* Return a pointer to the current `struct __locale_data' for CATEGORY.  */
 284 # define _NL_CURRENT_DATA(category) \
 285   (_NL_CURRENT_LOCALE->__locales[category])
 286
 287 /* Extract the current CATEGORY locale's string for ITEM.  */
 288 # define _NL_CURRENT(category, item) \
 289   (_NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].string)
 290
 291 /* Extract the current CATEGORY locale's string for ITEM.  */
 292 # define _NL_CURRENT_WSTR(category, item) \
 293   ((wchar_t *) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].wstr)
 294
 295 /* Extract the current CATEGORY locale's word for ITEM.  */
 296 # define _NL_CURRENT_WORD(category, item) \
 297   ((uint32_t) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].word)
 298
 299 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 300 # define _NL_CURRENT_DEFINE(category) \
 301   /* No per-category variable here. */
 302
 303 #endif
 304
 305 /* Extract CATEGORY locale's string for ITEM.  */
 306 static inline const char *
 307 _nl_lookup (locale_t l, int category, int item)
 308 {
 309   return l->__locales[category]->values[_NL_ITEM_INDEX (item)].string;
 310 }
 311
 312 /* Extract CATEGORY locale's wide string for ITEM.  */
 313 static inline const wchar_t *
 314 _nl_lookup_wstr (locale_t l, int category, int item)
 315 {
 316   return (wchar_t *) l->__locales[category]
 317     ->values[_NL_ITEM_INDEX (item)].wstr;
 318 }
 319
 320 /* Extract the CATEGORY locale's word for ITEM.  */
 321 static inline uint32_t
 322 _nl_lookup_word (locale_t l, int category, int item)
 323 {
 324   return l->__locales[category]->values[_NL_ITEM_INDEX (item)].word;
 325 }
 326
 327 /* Default search path if no LOCPATH environment variable.  */
 328 extern const char _nl_default_locale_path[] attribute_hidden;
 329
 330 /* Load the locale data for CATEGORY from the file specified by *NAME.
 331    If *NAME is "", use environment variables as specified by POSIX, and
 332    fill in *NAME with the actual name used.  If LOCALE_PATH is not null,
 333    those directories are searched for the locale files.  If it's null,
 334    the locale archive is checked first and then _nl_default_locale_path
 335    is searched for locale files.  */
 336 extern struct __locale_data *_nl_find_locale (const char *locale_path,
 337                                               size_t locale_path_len,
 338                                               int category, const char **name)
 339      internal_function attribute_hidden;
 340
 341 /* Try to load the file described by FILE.  */
 342 extern void _nl_load_locale (struct loaded_l10nfile *file, int category)
 343      internal_function attribute_hidden;
 344
 345 /* Free all resource.  */
 346 extern void _nl_unload_locale (struct __locale_data *locale)
 347      internal_function attribute_hidden;
 348
 349 /* Free the locale and give back all memory if the usage count is one.  */
 350 extern void _nl_remove_locale (int locale, struct __locale_data *data)
 351      internal_function attribute_hidden;
 352
 353 /* Find the locale *NAMEP in the locale archive, and return the
 354    internalized data structure for its CATEGORY data.  If this locale has
 355    already been loaded from the archive, just returns the existing data
 356    structure.  If successful, sets *NAMEP to point directly into the mapped
 357    archive string table; that way, the next call can short-circuit strcmp.  */
 358 extern struct __locale_data *_nl_load_locale_from_archive (int category,
 359                                                            const char **namep)
 360      internal_function attribute_hidden;
 361
 362 /* Subroutine of setlocale's __libc_subfreeres hook.  */
 363 extern void _nl_archive_subfreeres (void) attribute_hidden;
 364
 365 /* Subroutine of gconv-db's __libc_subfreeres hook.  */
 366 extern void _nl_locale_subfreeres (void) attribute_hidden;
 367
 368 /* Validate the contents of a locale file and set up the in-core
 369    data structure to point into the data.  This leaves the `alloc'
 370    and `name' fields uninitialized, for the caller to fill in.
 371    If any bogons are detected in the data, this will refuse to
 372    intern it, and return a null pointer instead.  */
 373 extern struct __locale_data *_nl_intern_locale_data (int category,
 374                                                      const void *data,
 375                                                      size_t datasize)
 376      internal_function attribute_hidden;
 377
 378
 379 /* Return `era' entry which corresponds to TP.  Used in strftime.  */
 380 extern struct era_entry *_nl_get_era_entry (const struct tm *tp,
 381                                             struct __locale_data *lc_time)
 382      internal_function attribute_hidden;
 383
 384 /* Return `era' cnt'th entry .  Used in strptime.  */
 385 extern struct era_entry *_nl_select_era_entry (int cnt,
 386                                                struct __locale_data *lc_time)
 387           internal_function attribute_hidden;
 388
 389 /* Return `alt_digit' which corresponds to NUMBER.  Used in strftime.  */
 390 extern const char *_nl_get_alt_digit (unsigned int number,
 391                                       struct __locale_data *lc_time)
 392           internal_function attribute_hidden;
 393
 394 /* Similar, but now for wide characters.  */
 395 extern const wchar_t *_nl_get_walt_digit (unsigned int number,
 396                                           struct __locale_data *lc_time)
 397      internal_function attribute_hidden;
 398
 399 /* Parse string as alternative digit and return numeric value.  */
 400 extern int _nl_parse_alt_digit (const char **strp,
 401                                 struct __locale_data *lc_time)
 402      internal_function attribute_hidden;
 403
 404 /* Postload processing.  */
 405 extern void _nl_postload_ctype (void);
 406
 407 /* Functions used for the `private.cleanup' hook.  */
 408 extern void _nl_cleanup_time (struct __locale_data *)
 409      internal_function attribute_hidden;
 410
 411
 412 #endif  /* localeinfo.h */