iconv/gconv_open.c

   1 /* Find matching transformation algorithms and initialize steps.
   2    Copyright (C) 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
   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 #include <errno.h>
  22 #include <locale.h>
  23 #include "../locale/localeinfo.h"
  24 #include <stdlib.h>
  25 #include <string.h>
  26
  27 #include <gconv_int.h>
  28
  29
  30 int
  31 internal_function
  32 __gconv_open (const char *toset, const char *fromset, __gconv_t *handle,
  33               int flags)
  34 {
  35   struct __gconv_step *steps;
  36   size_t nsteps;
  37   __gconv_t result = NULL;
  38   size_t cnt = 0;
  39   int res;
  40   int conv_flags = 0;
  41   const char *errhand;
  42   const char *ignore;
  43   struct trans_struct *trans = NULL;
  44
  45   /* Find out whether any error handling method is specified.  */
  46   errhand = strchr (toset, '/');
  47   if (errhand != NULL)
  48     errhand = strchr (errhand + 1, '/');
  49   if (__builtin_expect (errhand != NULL, 1))
  50     {
  51       if (*++errhand == '\0')
  52         errhand = NULL;
  53       else
  54         {
  55           /* Make copy without the error handling description.  */
  56           char *newtoset = (char *) alloca (errhand - toset + 1);
  57           char *tok;
  58           char *ptr;
  59
  60           newtoset[errhand - toset] = '\0';
  61           toset = memcpy (newtoset, toset, errhand - toset);
  62
  63           /* Find the appropriate transliteration handlers.  */
  64           tok = strdupa (errhand);
  65
  66           tok = __strtok_r (tok, ",", &ptr);
  67           while (tok != NULL)
  68             {
  69               if (__strcasecmp_l (tok, "TRANSLIT", &_nl_C_locobj) == 0)
  70                 {
  71                   /* It's the builtin transliteration handling.  We only
  72                      support it for working on the internal encoding.  */
  73                   static const char *internal_trans_names[1] = { "INTERNAL" };
  74                   struct trans_struct *lastp = NULL;
  75                   struct trans_struct *runp;
  76
  77                   for (runp = trans; runp != NULL; runp = runp->next)
  78                     if (runp->trans_fct == __gconv_transliterate)
  79                       break;
  80                     else
  81                       lastp = runp;
  82
  83                   if (runp == NULL)
  84                     {
  85                       struct trans_struct *newp;
  86
  87                       newp = (struct trans_struct *) alloca (sizeof (*newp));
  88                       memset (newp, '\0', sizeof (*newp));
  89
  90                       /* We leave the `name' field zero to signal that
  91                          this is an internal transliteration step.  */
  92                       newp->csnames = internal_trans_names;
  93                       newp->ncsnames = 1;
  94                       newp->trans_fct = __gconv_transliterate;
  95
  96                       if (lastp == NULL)
  97                         trans = newp;
  98                       else
  99                         lastp->next = newp;
 100                     }
 101                 }
 102               else if (__strcasecmp_l (tok, "IGNORE", &_nl_C_locobj) == 0)
 103                 /* Set the flag to ignore all errors.  */
 104                 conv_flags |= __GCONV_IGNORE_ERRORS;
 105               else
 106                 {
 107                   /* `tok' is possibly a module name.  We'll see later
 108                      whether we can find it.  But first see that we do
 109                      not already a module of this name.  */
 110                   struct trans_struct *lastp = NULL;
 111                   struct trans_struct *runp;
 112
 113                   for (runp = trans; runp != NULL; runp = runp->next)
 114                     if (runp->name != NULL
 115                         && __strcasecmp_l (tok, runp->name,
 116                                            &_nl_C_locobj) == 0)
 117                       break;
 118                     else
 119                       lastp = runp;
 120
 121                   if (runp == NULL)
 122                     {
 123                       struct trans_struct *newp;
 124
 125                       newp = (struct trans_struct *) alloca (sizeof (*newp));
 126                       memset (newp, '\0', sizeof (*newp));
 127                       newp->name = tok;
 128
 129                       if (lastp == NULL)
 130                         trans = newp;
 131                       else
 132                         lastp->next = newp;
 133                     }
 134                 }
 135
 136               tok = __strtok_r (NULL, ",", &ptr);
 137             }
 138         }
 139     }
 140
 141   /* For the source character set we ignore the error handler specification.
 142      XXX Is this really always the best?  */
 143   ignore = strchr (fromset, '/');
 144   if (ignore != NULL && (ignore = strchr (ignore + 1, '/')) != NULL
 145       && *++ignore != '\0')
 146     {
 147       char *newfromset = (char *) alloca (ignore - fromset + 1);
 148
 149       newfromset[ignore - fromset] = '\0';
 150       fromset = memcpy (newfromset, fromset, ignore - fromset);
 151     }
 152
 153   /* If the string is empty define this to mean the charset of the
 154      currently selected locale.  */
 155   if (strcmp (toset, "//") == 0)
 156     {
 157       const char *codeset = _NL_CURRENT (LC_CTYPE, CODESET);
 158       size_t len = strlen (codeset);
 159       char *dest;
 160       toset = dest = (char *) alloca (len + 3);
 161       memcpy (__mempcpy (dest, codeset, len), "//", 3);
 162     }
 163   if (strcmp (fromset, "//") == 0)
 164     {
 165       const char *codeset = _NL_CURRENT (LC_CTYPE, CODESET);
 166       size_t len = strlen (codeset);
 167       char *dest;
 168       fromset = dest = (char *) alloca (len + 3);
 169       memcpy (__mempcpy (dest, codeset, len), "//", 3);
 170     }
 171
 172   res = __gconv_find_transform (toset, fromset, &steps, &nsteps, flags);
 173   if (res == __GCONV_OK)
 174     {
 175       /* Find the modules.  */
 176       struct trans_struct *lastp = NULL;
 177       struct trans_struct *runp;
 178
 179       for (runp = trans; runp != NULL; runp = runp->next)
 180         {
 181           if (runp->name == NULL
 182               || __builtin_expect (__gconv_translit_find (runp), 0) == 0)
 183             lastp = runp;
 184           else
 185             /* This means we haven't found the module.  Remove it.  */
 186             (lastp == NULL ? trans : lastp->next) = runp->next;
 187         }
 188
 189       /* Allocate room for handle.  */
 190       result = (__gconv_t) malloc (sizeof (struct __gconv_info)
 191                                    + (nsteps
 192                                       * sizeof (struct __gconv_step_data)));
 193       if (result == NULL)
 194         res = __GCONV_NOMEM;
 195       else
 196         {
 197           size_t n;
 198
 199           /* Remember the list of steps.  */
 200           result->__steps = steps;
 201           result->__nsteps = nsteps;
 202
 203           /* Clear the array for the step data.  */
 204           memset (result->__data, '\0',
 205                   nsteps * sizeof (struct __gconv_step_data));
 206
 207           /* Call all initialization functions for the transformation
 208              step implementations.  */
 209           for (cnt = 0; cnt < nsteps; ++cnt)
 210             {
 211               size_t size;
 212
 213               /* Would have to be done if we would not clear the whole
 214                  array above.  */
 215 #if 0
 216               /* Reset the counter.  */
 217               result->__data[cnt].__invocation_counter = 0;
 218
 219               /* It's a regular use.  */
 220               result->__data[cnt].__internal_use = 0;
 221 #endif
 222
 223               /* We use the `mbstate_t' member in DATA.  */
 224               result->__data[cnt].__statep = &result->__data[cnt].__state;
 225
 226               /* Now see whether we can use any of the transliteration
 227                  modules for this step.  */
 228               for (runp = trans; runp != NULL; runp = runp->next)
 229                 for (n = 0; n < runp->ncsnames; ++n)
 230                   if (__strcasecmp_l (steps[cnt].__from_name,
 231                                       runp->csnames[n], &_nl_C_locobj) == 0)
 232                     {
 233                       void *data = NULL;
 234
 235                       /* Match!  Now try the initializer.  */
 236                       if (runp->trans_init_fct == NULL
 237                           || (runp->trans_init_fct (&data,
 238                                                     steps[cnt].__to_name)
 239                               == __GCONV_OK))
 240                         {
 241                           /* Append at the end of the list.  */
 242                           struct __gconv_trans_data *newp;
 243                           struct __gconv_trans_data **lastp;
 244
 245                           newp = (struct __gconv_trans_data *)
 246                             malloc (sizeof (struct __gconv_trans_data));
 247                           if (newp == NULL)
 248                             {
 249                               res = __GCONV_NOMEM;
 250                               goto bail;
 251                             }
 252
 253                           newp->__trans_fct = runp->trans_fct;
 254                           newp->__trans_context_fct = runp->trans_context_fct;
 255                           newp->__trans_end_fct = runp->trans_end_fct;
 256                           newp->__data = data;
 257                           newp->__next = NULL;
 258
 259                           lastp = &result->__data[cnt].__trans;
 260                           while (*lastp != NULL)
 261                             lastp = &(*lastp)->__next;
 262
 263                           *lastp = newp;
 264                         }
 265                       break;
 266                     }
 267
 268               /* If this is the last step we must not allocate an
 269                  output buffer.  */
 270               if (cnt < nsteps - 1)
 271                 {
 272                   result->__data[cnt].__flags = conv_flags;
 273
 274                   /* Allocate the buffer.  */
 275                   size = (GCONV_NCHAR_GOAL * steps[cnt].__max_needed_to);
 276
 277                   result->__data[cnt].__outbuf = (char *) malloc (size);
 278                   if (result->__data[cnt].__outbuf == NULL)
 279                     {
 280                       res = __GCONV_NOMEM;
 281                       goto bail;
 282                     }
 283
 284                   result->__data[cnt].__outbufend =
 285                     result->__data[cnt].__outbuf + size;
 286                 }
 287               else
 288                 {
 289                   /* Handle the last entry.  */
 290                   result->__data[cnt].__flags = conv_flags | __GCONV_IS_LAST;
 291
 292                   break;
 293                 }
 294             }
 295         }
 296
 297       if (res != __GCONV_OK)
 298         {
 299           /* Something went wrong.  Free all the resources.  */
 300           int serrno;
 301         bail:
 302           serrno = errno;
 303
 304           if (result != NULL)
 305             {
 306               while (cnt-- > 0)
 307                 {
 308                   struct __gconv_trans_data *transp;
 309
 310                   transp = result->__data[cnt].__trans;
 311                   while (transp != NULL)
 312                     {
 313                       struct __gconv_trans_data *curp = transp;
 314                       transp = transp->__next;
 315
 316                       if (__builtin_expect (curp->__trans_end_fct != NULL, 0))
 317                         curp->__trans_end_fct (curp->__data);
 318
 319                       free (curp);
 320                     }
 321
 322                   free (result->__data[cnt].__outbuf);
 323                 }
 324
 325               free (result);
 326               result = NULL;
 327             }
 328
 329           __gconv_close_transform (steps, nsteps);
 330
 331           __set_errno (serrno);
 332         }
 333     }
 334
 335   *handle = result;
 336   return res;
 337 }