iconv/gconv_int.h

   1 /* Copyright (C) 1997-2024 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3
   4    The GNU C Library is free software; you can redistribute it and/or
   5    modify it under the terms of the GNU Lesser General Public
   6    License as published by the Free Software Foundation; either
   7    version 2.1 of the License, or (at your option) any later version.
   8
   9    The GNU C Library is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12    Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser General Public
  15    License along with the GNU C Library; if not, see
  16    <https://www.gnu.org/licenses/>.  */
  17
  18 #ifndef _GCONV_INT_H
  19 #define _GCONV_INT_H    1
  20
  21 #include "gconv.h"
  22 #include <stdlib.h>             /* For alloca used in macro below.  */
  23 #include <ctype.h>              /* For __toupper_l used in macro below.  */
  24 #include <string.h>             /* For strlen et al used in macro below.  */
  25 #include <libc-lock.h>
  26
  27 __BEGIN_DECLS
  28
  29 /* We have to provide support for machines which are not able to handled
  30    unaligned memory accesses.  Some of the character encodings have
  31    representations with a fixed width of 2 or 4 bytes.  */
  32 #define get16(addr)                                                     \
  33 ({                                                                      \
  34   const struct { uint16_t r; } __attribute__ ((__packed__)) *__ptr      \
  35     = (__typeof(__ptr))(addr);                                          \
  36   __ptr->r;                                                             \
  37 })
  38 #define get32(addr)                                                     \
  39 ({                                                                      \
  40   const struct { uint32_t r; } __attribute__ ((__packed__)) *__ptr      \
  41     = (__typeof(__ptr))(addr);                                          \
  42   __ptr->r;                                                             \
  43 })
  44
  45 #define put16(addr, val)                                                \
  46 do {                                                                    \
  47    struct { uint16_t r; } __attribute__ ((__packed__)) *__ptr           \
  48     = (__typeof(__ptr))(addr);                                          \
  49    __ptr->r = val;                                                      \
  50 } while (0)
  51 #define put32(addr, val)                                                \
  52 do {                                                                    \
  53    struct { uint32_t r; } __attribute__ ((__packed__)) *__ptr           \
  54     = (__typeof(__ptr))(addr);                                          \
  55    __ptr->r = val;                                                      \
  56 } while (0)
  57
  58 /* Structure for alias definition.  Simply two strings.  */
  59 struct gconv_alias
  60 {
  61   char *fromname;
  62   char *toname;
  63 };
  64
  65
  66 /* Structure describing one loaded shared object.  This normally are
  67    objects to perform conversation but as a special case the db shared
  68    object is also handled.  */
  69 struct __gconv_loaded_object
  70 {
  71   /* Name of the object.  It must be the first structure element.  */
  72   const char *name;
  73
  74   /* Reference counter for the db functionality.  If no conversion is
  75      needed we unload the db library.  */
  76   int counter;
  77
  78   /* The handle for the shared object.  */
  79   void *handle;
  80
  81   /* Pointer to the functions the module defines.  */
  82   __gconv_fct fct;
  83   __gconv_init_fct init_fct;
  84   __gconv_end_fct end_fct;
  85 };
  86
  87
  88 /* Description for an available conversion module.  */
  89 struct gconv_module
  90 {
  91   const char *from_string;
  92   const char *to_string;
  93
  94   int cost_hi;
  95   int cost_lo;
  96
  97   const char *module_name;
  98
  99   struct gconv_module *left;    /* Prefix smaller.  */
 100   struct gconv_module *same;    /* List of entries with identical prefix.  */
 101   struct gconv_module *right;   /* Prefix larger.  */
 102 };
 103
 104
 105 /* The specification of the conversion that needs to be performed.  */
 106 struct gconv_spec
 107 {
 108   char *fromcode;
 109   char *tocode;
 110   bool translit;
 111   bool ignore;
 112 };
 113
 114 /* Flags for `gconv_open'.  */
 115 enum
 116 {
 117   GCONV_AVOID_NOCONV = 1 << 0
 118 };
 119
 120 /* When GCONV_AVOID_NOCONV is set and no conversion is needed,
 121    __GCONV_NULCONV should be returned.  */
 122 enum
 123 {
 124   __GCONV_NULCONV = -1
 125 };
 126
 127 /* Global variables.  */
 128
 129 /* Database of alias names.  */
 130 extern void *__gconv_alias_db attribute_hidden;
 131
 132 /* Array with available modules.  */
 133 extern struct gconv_module *__gconv_modules_db attribute_hidden;
 134
 135 /* Value of the GCONV_PATH environment variable.  */
 136 extern const char *__gconv_path_envvar attribute_hidden;
 137
 138 /* Lock for the conversion database content.  */
 139 __libc_lock_define (extern, __gconv_lock attribute_hidden)
 140
 141
 142 /* The gconv functions expects the name to be in upper case and complete,
 143    including the trailing slashes if necessary.  */
 144 #define norm_add_slashes(str,suffix) \
 145   ({                                                                          \
 146     const char *cp = (str);                                                   \
 147     char *result;                                                             \
 148     char *tmp;                                                                \
 149     size_t cnt = 0;                                                           \
 150     const size_t suffix_len = strlen (suffix);                                \
 151                                                                               \
 152     while (*cp != '\0')                                                       \
 153       if (*cp++ == '/')                                                       \
 154         ++cnt;                                                                \
 155                                                                               \
 156     tmp = result = __alloca (cp - (str) + 3 + suffix_len);                    \
 157     cp = (str);                                                               \
 158     while (*cp != '\0')                                                       \
 159       *tmp++ = __toupper_l (*cp++, _nl_C_locobj_ptr);                         \
 160     if (cnt < 2)                                                              \
 161       {                                                                       \
 162         *tmp++ = '/';                                                         \
 163         if (cnt < 1)                                                          \
 164           {                                                                   \
 165             *tmp++ = '/';                                                     \
 166             if (suffix_len != 0)                                              \
 167               tmp = __mempcpy (tmp, suffix, suffix_len);                      \
 168           }                                                                   \
 169       }                                                                       \
 170     *tmp = '\0';                                                              \
 171     result;                                                                   \
 172   })
 173
 174
 175 /* Return in *HANDLE, a descriptor for the transformation.  The function expects
 176    the specification of the transformation in the structure pointed to by
 177    CONV_SPEC.  It only reads *CONV_SPEC and does not take ownership of it.  */
 178 extern int __gconv_open (struct gconv_spec *conv_spec,
 179                          __gconv_t *handle, int flags);
 180 libc_hidden_proto (__gconv_open)
 181
 182 /* This function accepts the charset names of the source and destination of the
 183    conversion and populates *conv_spec with an equivalent conversion
 184    specification that may later be used by __gconv_open.  The charset names
 185    might contain options in the form of suffixes that alter the conversion,
 186    e.g. "ISO-10646/UTF-8/TRANSLIT".  It processes the charset names, ignoring
 187    and truncating any suffix options in fromcode, and processing and truncating
 188    any suffix options in tocode.  Supported suffix options ("TRANSLIT" or
 189    "IGNORE") when found in tocode lead to the corresponding flag in *conv_spec
 190    to be set to true.  Unrecognized suffix options are silently discarded.  If
 191    the function succeeds, it returns conv_spec back to the caller.  It returns
 192    NULL upon failure.  */
 193 extern struct gconv_spec *
 194 __gconv_create_spec (struct gconv_spec *conv_spec, const char *fromcode,
 195                      const char *tocode);
 196 libc_hidden_proto (__gconv_create_spec)
 197
 198 /* This function frees all heap memory allocated by __gconv_create_spec.  */
 199 extern void
 200 __gconv_destroy_spec (struct gconv_spec *conv_spec);
 201 libc_hidden_proto (__gconv_destroy_spec)
 202
 203 /* Free resources associated with transformation descriptor CD.  */
 204 extern int __gconv_close (__gconv_t cd)
 205      attribute_hidden;
 206
 207 /* Transform at most *INBYTESLEFT bytes from buffer starting at *INBUF
 208    according to rules described by CD and place up to *OUTBYTESLEFT
 209    bytes in buffer starting at *OUTBUF.  Return number of non-identical
 210    conversions in *IRREVERSIBLE if this pointer is not null.  */
 211 extern int __gconv (__gconv_t cd, const unsigned char **inbuf,
 212                     const unsigned char *inbufend, unsigned char **outbuf,
 213                     unsigned char *outbufend, size_t *irreversible)
 214      attribute_hidden;
 215
 216 /* Return in *HANDLE a pointer to an array with *NSTEPS elements describing
 217    the single steps necessary for transformation from FROMSET to TOSET.  */
 218 extern int __gconv_find_transform (const char *toset, const char *fromset,
 219                                    struct __gconv_step **handle,
 220                                    size_t *nsteps, int flags)
 221      attribute_hidden;
 222
 223 /* Search for transformation in cache data.  */
 224 extern int __gconv_lookup_cache (const char *toset, const char *fromset,
 225                                  struct __gconv_step **handle, size_t *nsteps,
 226                                  int flags)
 227      attribute_hidden;
 228
 229 /* Compare the two name for whether they are after alias expansion the
 230    same.  This function uses the cache and fails if none is
 231    loaded.  */
 232 extern int __gconv_compare_alias_cache (const char *name1, const char *name2,
 233                                         int *result)
 234      attribute_hidden;
 235
 236 /* Free data associated with a step's structure.  */
 237 extern void __gconv_release_step (struct __gconv_step *step)
 238      attribute_hidden;
 239
 240 /* Read all the configuration data and cache it if not done so already.  */
 241 extern void __gconv_load_conf (void) attribute_hidden;
 242
 243 /* Try to read module cache file.  */
 244 extern int __gconv_load_cache (void) attribute_hidden;
 245
 246 /* Retrieve pointer to internal cache.  */
 247 extern void *__gconv_get_cache (void);
 248
 249 /* Retrieve pointer to internal module database.  */
 250 extern struct gconv_module *__gconv_get_modules_db (void);
 251
 252 /* Retrieve pointer to internal alias database.  */
 253 extern void *__gconv_get_alias_db (void);
 254
 255 /* Comparison function to search alias.  */
 256 extern int __gconv_alias_compare (const void *p1, const void *p2)
 257      attribute_hidden;
 258
 259 /* Clear reference to transformation step implementations which might
 260    cause the code to be unloaded.  */
 261 extern int __gconv_close_transform (struct __gconv_step *steps,
 262                                     size_t nsteps)
 263      attribute_hidden;
 264
 265 /* Free all resources allocated for the transformation record when
 266    using the cache.  */
 267 extern void __gconv_release_cache (struct __gconv_step *steps, size_t nsteps)
 268      attribute_hidden;
 269
 270 /* Load shared object named by NAME.  If already loaded increment reference
 271    count.  */
 272 extern struct __gconv_loaded_object *__gconv_find_shlib (const char *name)
 273      attribute_hidden;
 274
 275 /* Release shared object.  If no further reference is available unload
 276    the object.  */
 277 extern void __gconv_release_shlib (struct __gconv_loaded_object *handle)
 278      attribute_hidden;
 279
 280 /* Fill STEP with information about builtin module with NAME.  */
 281 extern void __gconv_get_builtin_trans (const char *name,
 282                                        struct __gconv_step *step)
 283      attribute_hidden;
 284
 285 /* Transliteration using the locale's data.  */
 286 extern int __gconv_transliterate (struct __gconv_step *step,
 287                                   struct __gconv_step_data *step_data,
 288                                   const unsigned char *inbufstart,
 289                                   const unsigned char **inbufp,
 290                                   const unsigned char *inbufend,
 291                                   unsigned char **outbufstart,
 292                                   size_t *irreversible);
 293 libc_hidden_proto (__gconv_transliterate)
 294
 295 /* If NAME is an codeset alias expand it.  */
 296 extern int __gconv_compare_alias (const char *name1, const char *name2)
 297      attribute_hidden;
 298
 299
 300 /* Builtin transformations.  */
 301 #ifdef _LIBC
 302 # define __BUILTIN_TRANSFORM(Name) \
 303   extern int Name (struct __gconv_step *step,                                 \
 304                    struct __gconv_step_data *data,                            \
 305                    const unsigned char **inbuf,                               \
 306                    const unsigned char *inbufend,                             \
 307                    unsigned char **outbufstart, size_t *irreversible,         \
 308                    int do_flush, int consume_incomplete)
 309
 310 __BUILTIN_TRANSFORM (__gconv_transform_ascii_internal);
 311 __BUILTIN_TRANSFORM (__gconv_transform_internal_ascii);
 312 __BUILTIN_TRANSFORM (__gconv_transform_utf8_internal);
 313 __BUILTIN_TRANSFORM (__gconv_transform_internal_utf8);
 314 __BUILTIN_TRANSFORM (__gconv_transform_ucs2_internal);
 315 __BUILTIN_TRANSFORM (__gconv_transform_internal_ucs2);
 316 __BUILTIN_TRANSFORM (__gconv_transform_ucs2reverse_internal);
 317 __BUILTIN_TRANSFORM (__gconv_transform_internal_ucs2reverse);
 318 __BUILTIN_TRANSFORM (__gconv_transform_internal_ucs4);
 319 __BUILTIN_TRANSFORM (__gconv_transform_ucs4_internal);
 320 __BUILTIN_TRANSFORM (__gconv_transform_internal_ucs4le);
 321 __BUILTIN_TRANSFORM (__gconv_transform_ucs4le_internal);
 322 __BUILTIN_TRANSFORM (__gconv_transform_internal_utf16);
 323 __BUILTIN_TRANSFORM (__gconv_transform_utf16_internal);
 324 # undef __BUITLIN_TRANSFORM
 325
 326 /* Specialized conversion function for a single byte to INTERNAL, recognizing
 327    only ASCII characters.  */
 328 extern wint_t __gconv_btwoc_ascii (struct __gconv_step *step, unsigned char c);
 329
 330 #endif
 331
 332 __END_DECLS
 333
 334 #endif /* gconv_int.h */