locale/localeinfo.h

   1 /* Declarations for internal libc locale interfaces
   2    Copyright (C) 1995-2003, 2005, 2006, 2007, 2008, 2009, 2010, 2011
   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    : category == LC_CTYPE                                               \
  39    ? ((unsigned int) (0x20090720 ^ (category)))                         \
  40    : ((unsigned int) (0x20031115 ^ (category))))
  41
  42 /* Two special weight constants for the collation data.  */
  43 #define IGNORE_CHAR     2
  44
  45 /* We use a special value for the usage counter in `__locale_data' to
  46    signal that this data must never be removed anymore.  */
  47 #define MAX_USAGE_COUNT (UINT_MAX - 1)
  48 #define UNDELETABLE     UINT_MAX
  49
  50 /* Structure describing locale data in core for a category.  */
  51 struct __locale_data
  52 {
  53   const char *name;
  54   const char *filedata;         /* Region mapping the file data.  */
  55   off_t filesize;               /* Size of the file (and the region).  */
  56   enum                          /* Flavor of storage used for those.  */
  57   {
  58     ld_malloced,                /* Both are malloc'd.  */
  59     ld_mapped,                  /* name is malloc'd, filedata mmap'd */
  60     ld_archive                  /* Both point into mmap'd archive regions.  */
  61   } alloc;
  62
  63   /* This provides a slot for category-specific code to cache data computed
  64      about this locale.  That code can set a cleanup function to deallocate
  65      the data.  */
  66   struct
  67   {
  68     void (*cleanup) (struct __locale_data *) internal_function;
  69     union
  70     {
  71       void *data;
  72       struct lc_time_data *time;
  73       const struct gconv_fcts *ctype;
  74     };
  75   } private;
  76
  77   unsigned int usage_count;     /* Counter for users.  */
  78
  79   int use_translit;             /* Nonzero if the mb*towv*() and wc*tomb()
  80                                    functions should use transliteration.  */
  81
  82   unsigned int nstrings;        /* Number of strings below.  */
  83   union locale_data_value
  84   {
  85     const uint32_t *wstr;
  86     const char *string;
  87     unsigned int word;          /* Note endian issues vs 64-bit pointers.  */
  88   }
  89   values __flexarr;     /* Items, usually pointers into `filedata'.  */
  90 };
  91
  92 /* We know three kinds of collation sorting rules.  */
  93 enum coll_sort_rule
  94 {
  95   illegal_0__,
  96   sort_forward,
  97   sort_backward,
  98   illegal_3__,
  99   sort_position,
 100   sort_forward_position,
 101   sort_backward_position,
 102   sort_mask
 103 };
 104
 105 /* We can map the types of the entries into a few categories.  */
 106 enum value_type
 107 {
 108   none,
 109   string,
 110   stringarray,
 111   byte,
 112   bytearray,
 113   word,
 114   stringlist,
 115   wordarray,
 116   wstring,
 117   wstringarray,
 118   wstringlist
 119 };
 120
 121
 122 /* Definitions for `era' information from LC_TIME.  */
 123 #define ERA_NAME_FORMAT_MEMBERS 4
 124 #define ERA_M_NAME   0
 125 #define ERA_M_FORMAT 1
 126 #define ERA_W_NAME   2
 127 #define ERA_W_FORMAT 3
 128
 129
 130 /* Structure to access `era' information from LC_TIME.  */
 131 struct era_entry
 132 {
 133   uint32_t direction;           /* Contains '+' or '-'.  */
 134   int32_t offset;
 135   int32_t start_date[3];
 136   int32_t stop_date[3];
 137   const char *era_name;
 138   const char *era_format;
 139   const wchar_t *era_wname;
 140   const wchar_t *era_wformat;
 141   int absolute_direction;
 142   /* absolute direction:
 143      +1 indicates that year number is higher in the future. (like A.D.)
 144      -1 indicates that year number is higher in the past. (like B.C.)  */
 145 };
 146
 147 /* Structure caching computed data about information from LC_TIME.
 148    The `private.time' member of `struct __locale_data' points to this.  */
 149 struct lc_time_data
 150 {
 151   struct era_entry *eras;
 152   size_t num_eras;
 153   int era_initialized;
 154
 155   const char **alt_digits;
 156   const wchar_t **walt_digits;
 157   int alt_digits_initialized;
 158   int walt_digits_initialized;
 159 };
 160
 161
 162 /* LC_CTYPE specific:
 163    Hardwired indices for standard wide character translation mappings.  */
 164 enum
 165 {
 166   __TOW_toupper = 0,
 167   __TOW_tolower = 1
 168 };
 169
 170
 171 /* LC_CTYPE specific:
 172    Access a wide character class with a single character index.
 173    _ISCTYPE (c, desc) = iswctype (btowc (c), desc).
 174    c must be an `unsigned char'.  desc must be a nonzero wctype_t.  */
 175 #define _ISCTYPE(c, desc) \
 176   (((((const uint32_t *) (desc)) - 8)[(c) >> 5] >> ((c) & 0x1f)) & 1)
 177
 178 /* Category name handling variables.  */
 179 #define CATNAMEMF(line) CATNAMEMF1 (line)
 180 #define CATNAMEMF1(line) str##line
 181 extern const union catnamestr_t
 182 {
 183   struct
 184   {
 185 #define DEFINE_CATEGORY(category, category_name, items, a) \
 186     char CATNAMEMF (__LINE__)[sizeof (category_name)];
 187 #include "categories.def"
 188 #undef DEFINE_CATEGORY
 189   };
 190   char str[0];
 191 } _nl_category_names attribute_hidden;
 192 extern const uint8_t _nl_category_name_idxs[__LC_LAST] attribute_hidden;
 193 extern const uint8_t _nl_category_name_sizes[__LC_LAST] attribute_hidden;
 194
 195 /* Name of the standard locales.  */
 196 extern const char _nl_C_name[] attribute_hidden;
 197 extern const char _nl_POSIX_name[] attribute_hidden;
 198
 199 /* The standard codeset.  */
 200 extern const char _nl_C_codeset[] attribute_hidden;
 201
 202 /* This is the internal locale_t object that holds the global locale
 203    controlled by calls to setlocale.  A thread's TSD locale pointer
 204    points to this when `uselocale (LC_GLOBAL_LOCALE)' is in effect.  */
 205 extern struct __locale_struct _nl_global_locale attribute_hidden;
 206
 207 /* This fetches the thread-local locale_t pointer, either one set with
 208    uselocale or &_nl_global_locale.  */
 209 #define _NL_CURRENT_LOCALE      (__libc_tsd_get (__locale_t, LOCALE))
 210 #include <bits/libc-tsd.h>
 211 __libc_tsd_define (extern, __locale_t, LOCALE)
 212
 213
 214 /* For static linking it is desireable to avoid always linking in the code
 215    and data for every category when we can tell at link time that they are
 216    unused.  We can manage this playing some tricks with weak references.
 217    But with thread-local locale settings, it becomes quite ungainly unless
 218    we can use __thread variables.  So only in that case do we attempt this.  */
 219 #if !defined SHARED && defined HAVE_WEAK_SYMBOLS
 220 # include <tls.h>
 221 # define NL_CURRENT_INDIRECT    1
 222 #endif
 223
 224 #ifdef NL_CURRENT_INDIRECT
 225
 226 /* For each category declare the thread-local variable for the current
 227    locale data.  This has an extra indirection so it points at the
 228    __locales[CATEGORY] element in either _nl_global_locale or the current
 229    locale object set by uselocale, which points at the actual data.  The
 230    reason for having these variables is so that references to particular
 231    categories will link in the lc-CATEGORY.c module to define this symbol,
 232    and we arrange that linking that module is what brings in all the code
 233    associated with this category.  */
 234 #define DEFINE_CATEGORY(category, category_name, items, a) \
 235 extern __thread struct __locale_data *const *_nl_current_##category \
 236   attribute_hidden attribute_tls_model_ie;
 237 #include "categories.def"
 238 #undef  DEFINE_CATEGORY
 239
 240 /* Return a pointer to the current `struct __locale_data' for CATEGORY.  */
 241 #define _NL_CURRENT_DATA(category)      (*_nl_current_##category)
 242
 243 /* Extract the current CATEGORY locale's string for ITEM.  */
 244 #define _NL_CURRENT(category, item) \
 245   ((*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].string)
 246
 247 /* Extract the current CATEGORY locale's string for ITEM.  */
 248 #define _NL_CURRENT_WSTR(category, item) \
 249   ((wchar_t *) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].wstr)
 250
 251 /* Extract the current CATEGORY locale's word for ITEM.  */
 252 #define _NL_CURRENT_WORD(category, item) \
 253   ((uint32_t) (*_nl_current_##category)->values[_NL_ITEM_INDEX (item)].word)
 254
 255 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 256 #define _NL_CURRENT_DEFINE(category) \
 257   __thread struct __locale_data *const *_nl_current_##category \
 258     attribute_hidden = &_nl_global_locale.__locales[category]; \
 259   asm (_NL_CURRENT_DEFINE_STRINGIFY (ASM_GLOBAL_DIRECTIVE) \
 260        " " __SYMBOL_PREFIX "_nl_current_" #category "_used\n" \
 261        _NL_CURRENT_DEFINE_ABS (_nl_current_##category##_used, 1));
 262 #define _NL_CURRENT_DEFINE_STRINGIFY(x) _NL_CURRENT_DEFINE_STRINGIFY_1 (x)
 263 #define _NL_CURRENT_DEFINE_STRINGIFY_1(x) #x
 264 #ifdef HAVE_ASM_SET_DIRECTIVE
 265 # define _NL_CURRENT_DEFINE_ABS(sym, val) ".set " #sym ", " #val
 266 #else
 267 # define _NL_CURRENT_DEFINE_ABS(sym, val) #sym " = " #val
 268 #endif
 269
 270 #else
 271
 272 /* All categories are always loaded in the shared library, so there is no
 273    point in having lots of separate symbols for linking.  */
 274
 275 /* Return a pointer to the current `struct __locale_data' for CATEGORY.  */
 276 # define _NL_CURRENT_DATA(category) \
 277   (_NL_CURRENT_LOCALE->__locales[category])
 278
 279 /* Extract the current CATEGORY locale's string for ITEM.  */
 280 # define _NL_CURRENT(category, item) \
 281   (_NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].string)
 282
 283 /* Extract the current CATEGORY locale's string for ITEM.  */
 284 # define _NL_CURRENT_WSTR(category, item) \
 285   ((wchar_t *) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].wstr)
 286
 287 /* Extract the current CATEGORY locale's word for ITEM.  */
 288 # define _NL_CURRENT_WORD(category, item) \
 289   ((uint32_t) _NL_CURRENT_DATA (category)->values[_NL_ITEM_INDEX (item)].word)
 290
 291 /* This is used in lc-CATEGORY.c to define _nl_current_CATEGORY.  */
 292 # define _NL_CURRENT_DEFINE(category) \
 293   /* No per-category variable here. */
 294
 295 #endif
 296
 297
 298 /* Default search path if no LOCPATH environment variable.  */
 299 extern const char _nl_default_locale_path[] attribute_hidden;
 300
 301 /* Load the locale data for CATEGORY from the file specified by *NAME.
 302    If *NAME is "", use environment variables as specified by POSIX, and
 303    fill in *NAME with the actual name used.  If LOCALE_PATH is not null,
 304    those directories are searched for the locale files.  If it's null,
 305    the locale archive is checked first and then _nl_default_locale_path
 306    is searched for locale files.  */
 307 extern struct __locale_data *_nl_find_locale (const char *locale_path,
 308                                               size_t locale_path_len,
 309                                               int category, const char **name)
 310      internal_function attribute_hidden;
 311
 312 /* Try to load the file described by FILE.  */
 313 extern void _nl_load_locale (struct loaded_l10nfile *file, int category)
 314      internal_function attribute_hidden;
 315
 316 /* Free all resource.  */
 317 extern void _nl_unload_locale (struct __locale_data *locale)
 318      internal_function attribute_hidden;
 319
 320 /* Free the locale and give back all memory if the usage count is one.  */
 321 extern void _nl_remove_locale (int locale, struct __locale_data *data)
 322      internal_function attribute_hidden;
 323
 324 /* Find the locale *NAMEP in the locale archive, and return the
 325    internalized data structure for its CATEGORY data.  If this locale has
 326    already been loaded from the archive, just returns the existing data
 327    structure.  If successful, sets *NAMEP to point directly into the mapped
 328    archive string table; that way, the next call can short-circuit strcmp.  */
 329 extern struct __locale_data *_nl_load_locale_from_archive (int category,
 330                                                            const char **namep)
 331      internal_function attribute_hidden;
 332
 333 /* Subroutine of setlocale's __libc_subfreeres hook.  */
 334 extern void _nl_archive_subfreeres (void) attribute_hidden;
 335
 336 /* Subroutine of gconv-db's __libc_subfreeres hook.  */
 337 extern void _nl_locale_subfreeres (void) attribute_hidden;
 338
 339 /* Validate the contents of a locale file and set up the in-core
 340    data structure to point into the data.  This leaves the `alloc'
 341    and `name' fields uninitialized, for the caller to fill in.
 342    If any bogons are detected in the data, this will refuse to
 343    intern it, and return a null pointer instead.  */
 344 extern struct __locale_data *_nl_intern_locale_data (int category,
 345                                                      const void *data,
 346                                                      size_t datasize)
 347      internal_function attribute_hidden;
 348
 349
 350 /* Return `era' entry which corresponds to TP.  Used in strftime.  */
 351 extern struct era_entry *_nl_get_era_entry (const struct tm *tp,
 352                                             struct __locale_data *lc_time)
 353      internal_function attribute_hidden;
 354
 355 /* Return `era' cnt'th entry .  Used in strptime.  */
 356 extern struct era_entry *_nl_select_era_entry (int cnt,
 357                                                struct __locale_data *lc_time)
 358           internal_function attribute_hidden;
 359
 360 /* Return `alt_digit' which corresponds to NUMBER.  Used in strftime.  */
 361 extern const char *_nl_get_alt_digit (unsigned int number,
 362                                       struct __locale_data *lc_time)
 363           internal_function attribute_hidden;
 364
 365 /* Similar, but now for wide characters.  */
 366 extern const wchar_t *_nl_get_walt_digit (unsigned int number,
 367                                           struct __locale_data *lc_time)
 368      internal_function attribute_hidden;
 369
 370 /* Parse string as alternative digit and return numeric value.  */
 371 extern int _nl_parse_alt_digit (const char **strp,
 372                                 struct __locale_data *lc_time)
 373      internal_function attribute_hidden;
 374
 375 /* Postload processing.  */
 376 extern void _nl_postload_ctype (void);
 377
 378 /* Functions used for the `private.cleanup' hook.  */
 379 extern void _nl_cleanup_time (struct __locale_data *)
 380      internal_function attribute_hidden;
 381
 382
 383 #endif  /* localeinfo.h */