locale/localeinfo.h

   1 /* Declarations for internal libc locale interfaces
   2    Copyright (C) 1995-2003, 2005, 2006, 2007, 2008
   3    Free Software Foundation, Inc.
   4    This file is part of the GNU C Library.
   5
   6    The GNU C Library is free software; you can redistribute it and/or
   7    modify it under the terms of the GNU Lesser General Public
   8    License as published by the Free Software Foundation; either
   9    version 2.1 of the License, or (at your option) any later version.
  10
  11    The GNU C Library is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14    Lesser General Public License for more details.
  15
  16    You should have received a copy of the GNU Lesser General Public
  17    License along with the GNU C Library; if not, write to the Free
  18    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  19    02111-1307 USA.  */
  20
  21 #ifndef _LOCALEINFO_H
  22 #define _LOCALEINFO_H 1
  23
  24 #include <stddef.h>
  25 #include <langinfo.h>
  26 #include <limits.h>
  27 #include <locale.h>
  28 #include <time.h>
  29 #include <stdint.h>
  30 #include <sys/types.h>
  31
  32 #include <intl/loadinfo.h>      /* For loaded_l10nfile definition.  */
  33
  34 /* Magic number at the beginning of a locale data file for CATEGORY.  */
  35 #define LIMAGIC(category) \
  36   (category == LC_COLLATE                                               \
  37    ? ((unsigned int) (0x20051014 ^ (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 /* We know three kinds of collation sorting rules.  */
  91 enum coll_sort_rule
  92 {
  93   illegal_0__,
  94   sort_forward,
  95   sort_backward,
  96   illegal_3__,
  97   sort_position,
  98   sort_forward_position,
  99   sort_backward_position,
 100   sort_mask
 101 };
 102
 103 /* We can map the types of the entries into a few categories.  */
 104 enum value_type
 105 {
 106   none,
 107   string,
 108   stringarray,
 109   byte,
 110   bytearray,
 111   word,
 112   stringlist,
 113   wordarray,
 114   wstring,
 115   wstringarray,
 116   wstringlist
 117 };
 118
 119
 120 /* Definitions for `era' information from LC_TIME.  */
 121 #define ERA_NAME_FORMAT_MEMBERS 4
 122 #define ERA_M_NAME   0
 123 #define ERA_M_FORMAT 1
 124 #define ERA_W_NAME   2
 125 #define ERA_W_FORMAT 3
 126
 127
 128 /* Structure to access `era' information from LC_TIME.  */
 129 struct era_entry
 130 {
 131   uint32_t direction;           /* Contains '+' or '-'.  */
 132   int32_t offset;
 133   int32_t start_date[3];
 134   int32_t stop_date[3];
 135   const char *era_name;
 136   const char *era_format;
 137   const wchar_t *era_wname;
 138   const wchar_t *era_wformat;
 139   int absolute_direction;
 140   /* absolute direction:
 141      +1 indicates that year number is higher in the future. (like A.D.)
 142      -1 indicates that year number is higher in the past. (like B.C.)  */
 143 };
 144
 145 /* Structure caching computed data about information from LC_TIME.
 146    The `private.time' member of `struct locale_data' points to this.  */
 147 struct lc_time_data
 148 {
 149   struct era_entry *eras;
 150   size_t num_eras;
 151   int era_initialized;
 152
 153   const char **alt_digits;
 154   const wchar_t **walt_digits;
 155   int alt_digits_initialized;
 156   int walt_digits_initialized;
 157 };
 158
 159
 160 /* LC_CTYPE specific:
 161    Hardwired indices for standard wide character translation mappings.  */
 162 enum
 163 {
 164   __TOW_toupper = 0,
 165   __TOW_tolower = 1
 166 };
 167
 168
 169 /* LC_CTYPE specific:
 170    Access a wide character class with a single character index.
 171    _ISCTYPE (c, desc) = iswctype (btowc (c), desc).
 172    c must be an `unsigned char'.  desc must be a nonzero wctype_t.  */
 173 #define _ISCTYPE(c, desc) \
 174   (((((const uint32_t *) (desc)) - 8)[(c) >> 5] >> ((c) & 0x1f)) & 1)
 175
 176 /* Category name handling variables.  */
 177 #define CATNAMEMF(line) CATNAMEMF1 (line)
 178 #define CATNAMEMF1(line) str##line
 179 extern const union catnamestr_t
 180 {
 181   struct
 182   {
 183 #define DEFINE_CATEGORY(category, category_name, items, a) \
 184     char CATNAMEMF (__LINE__)[sizeof (category_name)];
 185 #include "categories.def"
 186 #undef DEFINE_CATEGORY
 187   };
 188   char str[0];
 189 } _nl_category_names attribute_hidden;
 190 extern const uint8_t _nl_category_name_idxs[__LC_LAST] attribute_hidden;
 191 extern const uint8_t _nl_category_name_sizes[__LC_LAST] attribute_hidden;
 192
 193 /* Name of the standard locales.  */
 194 extern const char _nl_C_name[] attribute_hidden;
 195 extern const char _nl_POSIX_name[] attribute_hidden;
 196
 197 /* The standard codeset.  */
 198 extern const char _nl_C_codeset[] attribute_hidden;
 199
 200 /* This is the internal locale_t object that holds the global locale
 201    controlled by calls to setlocale.  A thread's TSD locale pointer
 202    points to this when `uselocale (LC_GLOBAL_LOCALE)' is in effect.  */
 203 extern struct __locale_struct _nl_global_locale attribute_hidden;
 204
 205 /* This fetches the thread-local locale_t pointer, either one set with
 206    uselocale or &_nl_global_locale.  */
 207 #define _NL_CURRENT_LOCALE      (__libc_tsd_get (__locale_t, LOCALE))
 208 #include <bits/libc-tsd.h>
 209 __libc_tsd_define (extern, __locale_t, LOCALE)
 210
 211
 212 /* For static linking it is desireable to avoid always linking in the code
 213    and data for every category when we can tell at link time that they are
 214    unused.  We can manage this playing some tricks with weak references.
 215    But with thread-local locale settings, it becomes quite ungainly unless
 216    we can use __thread variables.  So only in that case do we attempt this.  */
 217 #if !defined SHARED && defined HAVE___THREAD && defined HAVE_WEAK_SYMBOLS
 218 # include <tls.h>
 219 # define NL_CURRENT_INDIRECT    1
 220 #endif
 221
 222 #ifdef NL_CURRENT_INDIRECT
 223
 224 /* For each category declare the thread-local variable for the current
 225    locale data.  This has an extra indirection so it points at the
 226    __locales[CATEGORY] element in either _nl_global_locale or the current
 227    locale object set by uselocale, which points at the actual data.  The
 228    reason for having these variables is so that references to particular
 229    categories will link in the lc-CATEGORY.c module to define this symbol,
 230    and we arrange that linking that module is what brings in all the code
 231    associated with this category.  */
 232 #define DEFINE_CATEGORY(category, category_name, items, a) \
 233 extern __thread struct locale_data *const *_nl_current_##category \
 234   attribute_hidden attribute_tls_model_ie;
 235 #include "categories.def"
 236 #undef  DEFINE_CATEGORY
 237
 238 /* Return a pointer to the current `struct locale_data' for CATEGORY.  */
 239 #define _NL_CURRENT_DATA(category)      (*_nl_current_##category)
 240
 241 /* Extract the current CATEGORY locale's string for ITEM.  */
 242 #define _NL_CURRENT(category, item) \
 243   ((*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].string)
 244
 245 /* Extract the current CATEGORY locale's string for ITEM.  */
 246 #define _NL_CURRENT_WSTR(category, item) \
 247   ((wchar_t *) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].wstr)
 248
 249 /* Extract the current CATEGORY locale's word for ITEM.  */
 250 #define _NL_CURRENT_WORD(category, item) \
 251   ((uint32_t) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].word)
 252
 253 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 254 #define _NL_CURRENT_DEFINE(category) \
 255   __thread struct locale_data *const *_nl_current_##category \
 256     attribute_hidden = &_nl_global_locale.__locales[category]; \
 257   asm (_NL_CURRENT_DEFINE_STRINGIFY (ASM_GLOBAL_DIRECTIVE) \
 258        " " __SYMBOL_PREFIX "_nl_current_" #category "_used\n" \
 259        _NL_CURRENT_DEFINE_ABS (_nl_current_##category##_used, 1));
 260 #define _NL_CURRENT_DEFINE_STRINGIFY(x) _NL_CURRENT_DEFINE_STRINGIFY_1 (x)
 261 #define _NL_CURRENT_DEFINE_STRINGIFY_1(x) #x
 262 #ifdef HAVE_ASM_SET_DIRECTIVE
 263 # define _NL_CURRENT_DEFINE_ABS(sym, val) ".set " #sym ", " #val
 264 #else
 265 # define _NL_CURRENT_DEFINE_ABS(sym, val) #sym " = " #val
 266 #endif
 267
 268 #else
 269
 270 /* All categories are always loaded in the shared library, so there is no
 271    point in having lots of separate symbols for linking.  */
 272
 273 /* Return a pointer to the current `struct locale_data' for CATEGORY.  */
 274 # define _NL_CURRENT_DATA(category) \
 275   (_NL_CURRENT_LOCALE->__locales[category])
 276
 277 /* Extract the current CATEGORY locale's string for ITEM.  */
 278 # define _NL_CURRENT(category, item) \
 279   (_NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].string)
 280
 281 /* Extract the current CATEGORY locale's string for ITEM.  */
 282 # define _NL_CURRENT_WSTR(category, item) \
 283   ((wchar_t *) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].wstr)
 284
 285 /* Extract the current CATEGORY locale's word for ITEM.  */
 286 # define _NL_CURRENT_WORD(category, item) \
 287   ((uint32_t) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].word)
 288
 289 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 290 # define _NL_CURRENT_DEFINE(category) \
 291   /* No per-category variable here. */
 292
 293 #endif
 294
 295
 296 /* Default search path if no LOCPATH environment variable.  */
 297 extern const char _nl_default_locale_path[] attribute_hidden;
 298
 299 /* Load the locale data for CATEGORY from the file specified by *NAME.
 300    If *NAME is "", use environment variables as specified by POSIX, and
 301    fill in *NAME with the actual name used.  If LOCALE_PATH is not null,
 302    those directories are searched for the locale files.  If it's null,
 303    the locale archive is checked first and then _nl_default_locale_path
 304    is searched for locale files.  */
 305 extern struct locale_data *_nl_find_locale (const char *locale_path,
 306                                             size_t locale_path_len,
 307                                             int category, const char **name)
 308      internal_function attribute_hidden;
 309
 310 /* Try to load the file described by FILE.  */
 311 extern void _nl_load_locale (struct loaded_l10nfile *file, int category)
 312      internal_function attribute_hidden;
 313
 314 /* Free all resource.  */
 315 extern void _nl_unload_locale (struct locale_data *locale)
 316      internal_function attribute_hidden;
 317
 318 /* Free the locale and give back all memory if the usage count is one.  */
 319 extern void _nl_remove_locale (int locale, struct locale_data *data)
 320      internal_function attribute_hidden;
 321
 322 /* Find the locale *NAMEP in the locale archive, and return the
 323    internalized data structure for its CATEGORY data.  If this locale has
 324    already been loaded from the archive, just returns the existing data
 325    structure.  If successful, sets *NAMEP to point directly into the mapped
 326    archive string table; that way, the next call can short-circuit strcmp.  */
 327 extern struct locale_data *_nl_load_locale_from_archive (int category,
 328                                                          const char **namep)
 329      internal_function attribute_hidden;
 330
 331 /* Subroutine of setlocale's __libc_subfreeres hook.  */
 332 extern void _nl_archive_subfreeres (void) attribute_hidden;
 333
 334 /* Subroutine of gconv-db's __libc_subfreeres hook.  */
 335 extern void _nl_locale_subfreeres (void) attribute_hidden;
 336
 337 /* Validate the contents of a locale file and set up the in-core
 338    data structure to point into the data.  This leaves the `alloc'
 339    and `name' fields uninitialized, for the caller to fill in.
 340    If any bogons are detected in the data, this will refuse to
 341    intern it, and return a null pointer instead.  */
 342 extern struct locale_data *_nl_intern_locale_data (int category,
 343                                                    const void *data,
 344                                                    size_t datasize)
 345      internal_function attribute_hidden;
 346
 347
 348 /* Return `era' entry which corresponds to TP.  Used in strftime.  */
 349 extern struct era_entry *_nl_get_era_entry (const struct tm *tp,
 350                                             struct locale_data *lc_time)
 351      internal_function attribute_hidden;
 352
 353 /* Return `era' cnt'th entry .  Used in strptime.  */
 354 extern struct era_entry *_nl_select_era_entry (int cnt,
 355                                                struct locale_data *lc_time)
 356           internal_function attribute_hidden;
 357
 358 /* Return `alt_digit' which corresponds to NUMBER.  Used in strftime.  */
 359 extern const char *_nl_get_alt_digit (unsigned int number,
 360                                       struct locale_data *lc_time)
 361           internal_function attribute_hidden;
 362
 363 /* Similar, but now for wide characters.  */
 364 extern const wchar_t *_nl_get_walt_digit (unsigned int number,
 365                                           struct locale_data *lc_time)
 366      internal_function attribute_hidden;
 367
 368 /* Parse string as alternative digit and return numeric value.  */
 369 extern int _nl_parse_alt_digit (const char **strp,
 370                                 struct locale_data *lc_time)
 371      internal_function attribute_hidden;
 372
 373 /* Postload processing.  */
 374 extern void _nl_postload_ctype (void);
 375
 376 /* Functions used for the `private.cleanup' hook.  */
 377 extern void _nl_cleanup_time (struct locale_data *)
 378      internal_function attribute_hidden;
 379
 380
 381 #endif  /* localeinfo.h */