locale/localeinfo.h

   1 /* Declarations for internal libc locale interfaces
   2    Copyright (C) 1995-2003, 2005, 2006 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, write to the Free
  17    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  18    02111-1307 USA.  */
  19
  20 #ifndef _LOCALEINFO_H
  21 #define _LOCALEINFO_H 1
  22
  23 #include <stddef.h>
  24 #include <langinfo.h>
  25 #include <limits.h>
  26 #include <locale.h>
  27 #include <time.h>
  28 #include <stdint.h>
  29 #include <sys/types.h>
  30
  31 #include <intl/loadinfo.h>      /* For loaded_l10nfile definition.  */
  32
  33 /* Magic number at the beginning of a locale data file for CATEGORY.  */
  34 #define LIMAGIC(category) \
  35   (category == LC_COLLATE                                               \
  36    ? ((unsigned int) (0x20051014 ^ (category)))                         \
  37    : ((unsigned int) (0x20031115 ^ (category))))
  38
  39 /* Two special weight constants for the collation data.  */
  40 #define IGNORE_CHAR     2
  41
  42 /* We use a special value for the usage counter in `locale_data' to
  43    signal that this data must never be removed anymore.  */
  44 #define MAX_USAGE_COUNT (UINT_MAX - 1)
  45 #define UNDELETABLE     UINT_MAX
  46
  47 /* Structure describing locale data in core for a category.  */
  48 struct locale_data
  49 {
  50   const char *name;
  51   const char *filedata;         /* Region mapping the file data.  */
  52   off_t filesize;               /* Size of the file (and the region).  */
  53   enum                          /* Flavor of storage used for those.  */
  54   {
  55     ld_malloced,                /* Both are malloc'd.  */
  56     ld_mapped,                  /* name is malloc'd, filedata mmap'd */
  57     ld_archive                  /* Both point into mmap'd archive regions.  */
  58   } alloc;
  59
  60   /* This provides a slot for category-specific code to cache data computed
  61      about this locale.  That code can set a cleanup function to deallocate
  62      the data.  */
  63   struct
  64   {
  65     void (*cleanup) (struct locale_data *) internal_function;
  66     union
  67     {
  68       void *data;
  69       struct lc_time_data *time;
  70       const struct gconv_fcts *ctype;
  71     };
  72   } private;
  73
  74   unsigned int usage_count;     /* Counter for users.  */
  75
  76   int use_translit;             /* Nonzero if the mb*towv*() and wc*tomb()
  77                                    functions should use transliteration.  */
  78
  79   unsigned int nstrings;        /* Number of strings below.  */
  80   union locale_data_value
  81   {
  82     const uint32_t *wstr;
  83     const char *string;
  84     unsigned int word;          /* Note endian issues vs 64-bit pointers.  */
  85   }
  86   values __flexarr;     /* Items, usually pointers into `filedata'.  */
  87 };
  88
  89 /* We know three kinds of collation sorting rules.  */
  90 enum coll_sort_rule
  91 {
  92   illegal_0__,
  93   sort_forward,
  94   sort_backward,
  95   illegal_3__,
  96   sort_position,
  97   sort_forward_position,
  98   sort_backward_position,
  99   sort_mask
 100 };
 101
 102 /* We can map the types of the entries into a few categories.  */
 103 enum value_type
 104 {
 105   none,
 106   string,
 107   stringarray,
 108   byte,
 109   bytearray,
 110   word,
 111   stringlist,
 112   wordarray,
 113   wstring,
 114   wstringarray,
 115   wstringlist
 116 };
 117
 118
 119 /* Definitions for `era' information from LC_TIME.  */
 120 #define ERA_NAME_FORMAT_MEMBERS 4
 121 #define ERA_M_NAME   0
 122 #define ERA_M_FORMAT 1
 123 #define ERA_W_NAME   2
 124 #define ERA_W_FORMAT 3
 125
 126
 127 /* Structure to access `era' information from LC_TIME.  */
 128 struct era_entry
 129 {
 130   uint32_t direction;           /* Contains '+' or '-'.  */
 131   int32_t offset;
 132   int32_t start_date[3];
 133   int32_t stop_date[3];
 134   const char *era_name;
 135   const char *era_format;
 136   const wchar_t *era_wname;
 137   const wchar_t *era_wformat;
 138   int absolute_direction;
 139   /* absolute direction:
 140      +1 indicates that year number is higher in the future. (like A.D.)
 141      -1 indicates that year number is higher in the past. (like B.C.)  */
 142 };
 143
 144 /* Structure caching computed data about information from LC_TIME.
 145    The `private.time' member of `struct locale_data' points to this.  */
 146 struct lc_time_data
 147 {
 148   struct era_entry *eras;
 149   size_t num_eras;
 150   int era_initialized;
 151
 152   const char **alt_digits;
 153   const wchar_t **walt_digits;
 154   int alt_digits_initialized;
 155   int walt_digits_initialized;
 156 };
 157
 158
 159 /* LC_CTYPE specific:
 160    Hardwired indices for standard wide character translation mappings.  */
 161 enum
 162 {
 163   __TOW_toupper = 0,
 164   __TOW_tolower = 1
 165 };
 166
 167
 168 /* LC_CTYPE specific:
 169    Access a wide character class with a single character index.
 170    _ISCTYPE (c, desc) = iswctype (btowc (c), desc).
 171    c must be an `unsigned char'.  desc must be a nonzero wctype_t.  */
 172 #define _ISCTYPE(c, desc) \
 173   (((((const uint32_t *) (desc)) - 8)[(c) >> 5] >> ((c) & 0x1f)) & 1)
 174
 175 /* Category name handling variables.  */
 176 #define CATNAMEMF(line) CATNAMEMF1 (line)
 177 #define CATNAMEMF1(line) str##line
 178 extern const union catnamestr_t
 179 {
 180   struct
 181   {
 182 #define DEFINE_CATEGORY(category, category_name, items, a) \
 183     char CATNAMEMF (__LINE__)[sizeof (category_name)];
 184 #include "categories.def"
 185 #undef DEFINE_CATEGORY
 186   };
 187   char str[0];
 188 } _nl_category_names attribute_hidden;
 189 const uint8_t _nl_category_name_idxs[__LC_LAST] attribute_hidden;
 190 extern const uint8_t _nl_category_name_sizes[__LC_LAST] attribute_hidden;
 191
 192 /* Name of the standard locales.  */
 193 extern const char _nl_C_name[] attribute_hidden;
 194 extern const char _nl_POSIX_name[] attribute_hidden;
 195
 196 /* The standard codeset.  */
 197 extern const char _nl_C_codeset[] attribute_hidden;
 198
 199 /* This is the internal locale_t object that holds the global locale
 200    controlled by calls to setlocale.  A thread's TSD locale pointer
 201    points to this when `uselocale (LC_GLOBAL_LOCALE)' is in effect.  */
 202 extern struct __locale_struct _nl_global_locale attribute_hidden;
 203
 204 /* This fetches the thread-local locale_t pointer, either one set with
 205    uselocale or &_nl_global_locale.  */
 206 #define _NL_CURRENT_LOCALE      ((__locale_t) __libc_tsd_get (LOCALE))
 207 #include <bits/libc-tsd.h>
 208 __libc_tsd_define (extern, LOCALE)
 209
 210
 211 /* For static linking it is desireable to avoid always linking in the code
 212    and data for every category when we can tell at link time that they are
 213    unused.  We can manage this playing some tricks with weak references.
 214    But with thread-local locale settings, it becomes quite ungainly unless
 215    we can use __thread variables.  So only in that case do we attempt this.  */
 216 #if !defined SHARED && defined HAVE___THREAD && defined HAVE_WEAK_SYMBOLS
 217 # include <tls.h>
 218 # define NL_CURRENT_INDIRECT    1
 219 #endif
 220
 221 #ifdef NL_CURRENT_INDIRECT
 222
 223 /* For each category declare the thread-local variable for the current
 224    locale data.  This has an extra indirection so it points at the
 225    __locales[CATEGORY] element in either _nl_global_locale or the current
 226    locale object set by uselocale, which points at the actual data.  The
 227    reason for having these variables is so that references to particular
 228    categories will link in the lc-CATEGORY.c module to define this symbol,
 229    and we arrange that linking that module is what brings in all the code
 230    associated with this category.  */
 231 #define DEFINE_CATEGORY(category, category_name, items, a) \
 232 extern __thread struct locale_data *const *_nl_current_##category \
 233   attribute_hidden attribute_tls_model_ie;
 234 #include "categories.def"
 235 #undef  DEFINE_CATEGORY
 236
 237 /* Return a pointer to the current `struct locale_data' for CATEGORY.  */
 238 #define _NL_CURRENT_DATA(category)      (*_nl_current_##category)
 239
 240 /* Extract the current CATEGORY locale's string for ITEM.  */
 241 #define _NL_CURRENT(category, item) \
 242   ((*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].string)
 243
 244 /* Extract the current CATEGORY locale's string for ITEM.  */
 245 #define _NL_CURRENT_WSTR(category, item) \
 246   ((wchar_t *) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].wstr)
 247
 248 /* Extract the current CATEGORY locale's word for ITEM.  */
 249 #define _NL_CURRENT_WORD(category, item) \
 250   ((uint32_t) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].word)
 251
 252 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 253 #define _NL_CURRENT_DEFINE(category) \
 254   __thread struct locale_data *const *_nl_current_##category \
 255     attribute_hidden = &_nl_global_locale.__locales[category]; \
 256   asm (_NL_CURRENT_DEFINE_STRINGIFY (ASM_GLOBAL_DIRECTIVE) \
 257        " " __SYMBOL_PREFIX "_nl_current_" #category "_used\n" \
 258        _NL_CURRENT_DEFINE_ABS (_nl_current_##category##_used, 1));
 259 #define _NL_CURRENT_DEFINE_STRINGIFY(x) _NL_CURRENT_DEFINE_STRINGIFY_1 (x)
 260 #define _NL_CURRENT_DEFINE_STRINGIFY_1(x) #x
 261 #ifdef HAVE_ASM_SET_DIRECTIVE
 262 # define _NL_CURRENT_DEFINE_ABS(sym, val) ".set " #sym ", " #val
 263 #else
 264 # define _NL_CURRENT_DEFINE_ABS(sym, val) #sym " = " #val
 265 #endif
 266
 267 #else
 268
 269 /* All categories are always loaded in the shared library, so there is no
 270    point in having lots of separate symbols for linking.  */
 271
 272 /* Return a pointer to the current `struct locale_data' for CATEGORY.  */
 273 # define _NL_CURRENT_DATA(category) \
 274   (_NL_CURRENT_LOCALE->__locales[category])
 275
 276 /* Extract the current CATEGORY locale's string for ITEM.  */
 277 # define _NL_CURRENT(category, item) \
 278   (_NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].string)
 279
 280 /* Extract the current CATEGORY locale's string for ITEM.  */
 281 # define _NL_CURRENT_WSTR(category, item) \
 282   ((wchar_t *) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].wstr)
 283
 284 /* Extract the current CATEGORY locale's word for ITEM.  */
 285 # define _NL_CURRENT_WORD(category, item) \
 286   ((uint32_t) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].word)
 287
 288 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 289 # define _NL_CURRENT_DEFINE(category) \
 290   /* No per-category variable here. */
 291
 292 #endif
 293
 294
 295 /* Default search path if no LOCPATH environment variable.  */
 296 extern const char _nl_default_locale_path[] attribute_hidden;
 297
 298 /* Load the locale data for CATEGORY from the file specified by *NAME.
 299    If *NAME is "", use environment variables as specified by POSIX, and
 300    fill in *NAME with the actual name used.  If LOCALE_PATH is not null,
 301    those directories are searched for the locale files.  If it's null,
 302    the locale archive is checked first and then _nl_default_locale_path
 303    is searched for locale files.  */
 304 extern struct locale_data *_nl_find_locale (const char *locale_path,
 305                                             size_t locale_path_len,
 306                                             int category, const char **name)
 307      internal_function attribute_hidden;
 308
 309 /* Try to load the file described by FILE.  */
 310 extern void _nl_load_locale (struct loaded_l10nfile *file, int category)
 311      internal_function attribute_hidden;
 312
 313 /* Free all resource.  */
 314 extern void _nl_unload_locale (struct locale_data *locale)
 315      internal_function attribute_hidden;
 316
 317 /* Free the locale and give back all memory if the usage count is one.  */
 318 extern void _nl_remove_locale (int locale, struct locale_data *data)
 319      internal_function attribute_hidden;
 320
 321 /* Find the locale *NAMEP in the locale archive, and return the
 322    internalized data structure for its CATEGORY data.  If this locale has
 323    already been loaded from the archive, just returns the existing data
 324    structure.  If successful, sets *NAMEP to point directly into the mapped
 325    archive string table; that way, the next call can short-circuit strcmp.  */
 326 extern struct locale_data *_nl_load_locale_from_archive (int category,
 327                                                          const char **namep)
 328      internal_function attribute_hidden;
 329
 330 /* Subroutine of setlocale's __libc_subfreeres hook.  */
 331 extern void _nl_archive_subfreeres (void) attribute_hidden;
 332
 333 /* Subroutine of gconv-db's __libc_subfreeres hook.  */
 334 extern void _nl_locale_subfreeres (void) attribute_hidden;
 335
 336 /* Validate the contents of a locale file and set up the in-core
 337    data structure to point into the data.  This leaves the `alloc'
 338    and `name' fields uninitialized, for the caller to fill in.
 339    If any bogons are detected in the data, this will refuse to
 340    intern it, and return a null pointer instead.  */
 341 extern struct locale_data *_nl_intern_locale_data (int category,
 342                                                    const void *data,
 343                                                    size_t datasize)
 344      internal_function attribute_hidden;
 345
 346
 347 /* Return `era' entry which corresponds to TP.  Used in strftime.  */
 348 extern struct era_entry *_nl_get_era_entry (const struct tm *tp,
 349                                             struct locale_data *lc_time)
 350      internal_function attribute_hidden;
 351
 352 /* Return `era' cnt'th entry .  Used in strptime.  */
 353 extern struct era_entry *_nl_select_era_entry (int cnt,
 354                                                struct locale_data *lc_time)
 355           internal_function attribute_hidden;
 356
 357 /* Return `alt_digit' which corresponds to NUMBER.  Used in strftime.  */
 358 extern const char *_nl_get_alt_digit (unsigned int number,
 359                                       struct locale_data *lc_time)
 360           internal_function attribute_hidden;
 361
 362 /* Similar, but now for wide characters.  */
 363 extern const wchar_t *_nl_get_walt_digit (unsigned int number,
 364                                           struct locale_data *lc_time)
 365      internal_function attribute_hidden;
 366
 367 /* Parse string as alternative digit and return numeric value.  */
 368 extern int _nl_parse_alt_digit (const char **strp,
 369                                 struct locale_data *lc_time)
 370      internal_function attribute_hidden;
 371
 372 /* Postload processing.  */
 373 extern void _nl_postload_ctype (void);
 374
 375 /* Functions used for the `private.cleanup' hook.  */
 376 extern void _nl_cleanup_time (struct locale_data *)
 377      internal_function attribute_hidden;
 378
 379
 380 #endif  /* localeinfo.h */