iconv/gconv_conf.c

   1 /* Handle configuration data.
   2    Copyright (C) 1997, 1998, 1999, 2000 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 Library General Public License as
   8    published by the Free Software Foundation; either version 2 of the
   9    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    Library General Public License for more details.
  15
  16    You should have received a copy of the GNU Library General Public
  17    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  18    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19    Boston, MA 02111-1307, USA.  */
  20
  21 #include <assert.h>
  22 #include <ctype.h>
  23 #include <errno.h>
  24 #include <limits.h>
  25 #include <search.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <unistd.h>
  30 #include <sys/param.h>
  31
  32 #include <bits/libc-lock.h>
  33 #include <gconv_int.h>
  34
  35
  36 /* This is the default path where we look for module lists.  */
  37 static const char default_gconv_path[] = GCONV_PATH;
  38
  39 /* The path element in use.   */
  40 const struct path_elem *__gconv_path_elem;
  41 /* Maximum length of a single path element.  */
  42 size_t __gconv_max_path_elem_len;
  43
  44 /* We use the following struct if we couldn't allocate memory.  */
  45 static const struct path_elem empty_path_elem;
  46
  47 /* Name of the file containing the module information in the directories
  48    along the path.  */
  49 static const char gconv_conf_filename[] = "gconv-modules";
  50
  51 /* Filename extension for the modules.  */
  52 #ifndef MODULE_EXT
  53 # define MODULE_EXT ".so"
  54 #endif
  55 static const char gconv_module_ext[] = MODULE_EXT;
  56
  57 /* We have a few builtin transformations.  */
  58 static struct gconv_module builtin_modules[] =
  59 {
  60 #define BUILTIN_TRANSFORMATION(From, To, Cost, Name, Fct, Init, End, MinF, \
  61                                MaxF, MinT, MaxT) \
  62   {                                                                           \
  63     from_string: From,                                                        \
  64     to_string: To,                                                            \
  65     cost_hi: Cost,                                                            \
  66     cost_lo: INT_MAX,                                                         \
  67     module_name: Name                                                         \
  68   },
  69 #define BUILTIN_ALIAS(From, To)
  70
  71 #include "gconv_builtin.h"
  72 };
  73
  74 #undef BUILTIN_TRANSFORMATION
  75 #undef BUILTIN_ALIAS
  76
  77 static const char *builtin_aliases[] =
  78 {
  79 #define BUILTIN_TRANSFORMATION(From, To, Cost, Name, Fct, Init, End, MinF, \
  80                                MaxF, MinT, MaxT)
  81 #define BUILTIN_ALIAS(From, To) From " " To,
  82
  83 #include "gconv_builtin.h"
  84 };
  85
  86 #ifdef USE_IN_LIBIO
  87 # include <libio/libioP.h>
  88 # define __getdelim(line, len, c, fp) _IO_getdelim (line, len, c, fp)
  89 #endif
  90
  91
  92 /* Test whether there is already a matching module known.  */
  93 static int
  94 internal_function
  95 detect_conflict (const char *alias)
  96 {
  97   struct gconv_module *node = __gconv_modules_db;
  98
  99   while (node != NULL)
 100     {
 101       int cmpres = strcmp (alias, node->from_string);
 102
 103       if (cmpres == 0)
 104         /* We have a conflict.  */
 105         return 1;
 106       else if (cmpres < 0)
 107         node = node->left;
 108       else
 109         node = node->right;
 110     }
 111
 112   return node != NULL;
 113 }
 114
 115
 116 /* Add new alias.  */
 117 static inline void
 118 add_alias (char *rp, void *modules)
 119 {
 120   /* We now expect two more string.  The strings are normalized
 121      (converted to UPPER case) and strored in the alias database.  */
 122   struct gconv_alias *new_alias;
 123   char *from, *to, *wp;
 124
 125   while (isspace (*rp))
 126     ++rp;
 127   from = wp = rp;
 128   while (*rp != '\0' && !isspace (*rp))
 129     *wp++ = toupper (*rp++);
 130   if (*rp == '\0')
 131     /* There is no `to' string on the line.  Ignore it.  */
 132     return;
 133   *wp++ = '\0';
 134   to = ++rp;
 135   while (isspace (*rp))
 136     ++rp;
 137   while (*rp != '\0' && !isspace (*rp))
 138     *wp++ = toupper (*rp++);
 139   if (to == wp)
 140     /* No `to' string, ignore the line.  */
 141     return;
 142   *wp++ = '\0';
 143
 144   /* Test whether this alias conflicts with any available module.  */
 145   if (detect_conflict (from))
 146     /* It does conflict, don't add the alias.  */
 147     return;
 148
 149   new_alias = (struct gconv_alias *)
 150     malloc (sizeof (struct gconv_alias) + (wp - from));
 151   if (new_alias != NULL)
 152     {
 153       void **inserted;
 154
 155       new_alias->fromname = memcpy ((char *) new_alias
 156                                     + sizeof (struct gconv_alias),
 157                                     from, wp - from);
 158       new_alias->toname = new_alias->fromname + (to - from);
 159
 160       inserted = (void **) __tsearch (new_alias, &__gconv_alias_db,
 161                                       __gconv_alias_compare);
 162       if (inserted == NULL || *inserted != new_alias)
 163         /* Something went wrong, free this entry.  */
 164         free (new_alias);
 165     }
 166 }
 167
 168
 169 /* Insert a data structure for a new module in the search tree.  */
 170 static inline void
 171 internal_function
 172 insert_module (struct gconv_module *newp)
 173 {
 174   struct gconv_module **rootp = &__gconv_modules_db;
 175
 176   while (*rootp != NULL)
 177     {
 178       struct gconv_module *root = *rootp;
 179       int cmpres;
 180
 181       cmpres = strcmp (newp->from_string, root->from_string);
 182       if (cmpres == 0)
 183         {
 184           /* Both strings are identical.  Insert the string at the
 185              end of the `same' list if it is not already there.  */
 186           while (strcmp (newp->from_string, root->from_string) != 0
 187                  || strcmp (newp->to_string, root->to_string) != 0)
 188             {
 189               rootp = &root->same;
 190               root = *rootp;
 191               if (root == NULL)
 192                 break;
 193             }
 194
 195           if (root != NULL)
 196             /* This is a no new conversion.  */
 197             return;
 198
 199           break;
 200         }
 201       else if (cmpres < 0)
 202         rootp = &root->left;
 203       else
 204         rootp = &root->right;
 205     }
 206
 207   /* Plug in the new node here.  */
 208   *rootp = newp;
 209 }
 210
 211
 212 /* Add new module.  */
 213 static void
 214 internal_function
 215 add_module (char *rp, const char *directory, size_t dir_len, void **modules,
 216             size_t *nmodules, int modcounter)
 217 {
 218   /* We expect now
 219      1. `from' name
 220      2. `to' name
 221      3. filename of the module
 222      4. an optional cost value
 223   */
 224   struct gconv_alias fake_alias;
 225   struct gconv_module *new_module;
 226   char *from, *to, *module, *wp;
 227   int need_ext;
 228   int cost_hi;
 229
 230   while (isspace (*rp))
 231     ++rp;
 232   from = rp;
 233   while (*rp != '\0' && !isspace (*rp))
 234     {
 235       *rp = toupper (*rp);
 236       ++rp;
 237     }
 238   if (*rp == '\0')
 239     return;
 240   *rp++ = '\0';
 241   to = wp = rp;
 242   while (isspace (*rp))
 243     ++rp;
 244   while (*rp != '\0' && !isspace (*rp))
 245     *wp++ = toupper (*rp++);
 246   if (*rp == '\0')
 247     return;
 248   *wp++ = '\0';
 249   do
 250     ++rp;
 251   while (isspace (*rp));
 252   module = wp;
 253   while (*rp != '\0' && !isspace (*rp))
 254     *wp++ = *rp++;
 255   if (*rp == '\0')
 256     {
 257       /* There is no cost, use one by default.  */
 258       *wp++ = '\0';
 259       cost_hi = 1;
 260     }
 261   else
 262     {
 263       /* There might be a cost value.  */
 264       char *endp;
 265
 266       *wp++ = '\0';
 267       cost_hi = strtol (rp, &endp, 10);
 268       if (rp == endp || cost_hi < 1)
 269         /* No useful information.  */
 270         cost_hi = 1;
 271     }
 272
 273   if (module[0] == '\0')
 274     /* No module name given.  */
 275     return;
 276   if (module[0] == '/')
 277     dir_len = 0;
 278   else
 279     /* Increment by one for the slash.  */
 280     ++dir_len;
 281
 282   /* See whether we must add the ending.  */
 283   need_ext = 0;
 284   if (wp - module < sizeof (gconv_module_ext)
 285       || memcmp (wp - sizeof (gconv_module_ext), gconv_module_ext,
 286                  sizeof (gconv_module_ext)) != 0)
 287     /* We must add the module extension.  */
 288     need_ext = sizeof (gconv_module_ext) - 1;
 289
 290   /* See whether we have already an alias with this name defined.  */
 291   fake_alias.fromname = strndupa (from, to - from);
 292
 293   if (__tfind (&fake_alias, &__gconv_alias_db, __gconv_alias_compare) != NULL)
 294     /* This module duplicates an alias.  */
 295     return;
 296
 297   new_module = (struct gconv_module *) calloc (1,
 298                                                sizeof (struct gconv_module)
 299                                                + (wp - from)
 300                                                + dir_len + need_ext);
 301   if (new_module != NULL)
 302     {
 303       char *tmp;
 304
 305       new_module->from_string = memcpy ((char *) new_module
 306                                         + sizeof (struct gconv_module),
 307                                         from, to - from);
 308
 309       new_module->to_string = memcpy ((char *) new_module->from_string
 310                                       + (to - from), to, module - to);
 311
 312       new_module->cost_hi = cost_hi;
 313       new_module->cost_lo = modcounter;
 314
 315       new_module->module_name = (char *) new_module->to_string + (module - to);
 316
 317       if (dir_len == 0)
 318         tmp = (char *) new_module->module_name;
 319       else
 320         {
 321           tmp = __mempcpy ((char *) new_module->module_name,
 322                            directory, dir_len - 1);
 323           *tmp++ = '/';
 324         }
 325
 326       tmp = __mempcpy (tmp, module, wp - module);
 327
 328       if (need_ext)
 329         memcpy (tmp - 1, gconv_module_ext, sizeof (gconv_module_ext));
 330
 331       /* Now insert the new module data structure in our search tree.  */
 332       insert_module (new_module);
 333     }
 334 }
 335
 336
 337 /* Read the next configuration file.  */
 338 static void
 339 internal_function
 340 read_conf_file (const char *filename, const char *directory, size_t dir_len,
 341                 void **modules, size_t *nmodules)
 342 {
 343   FILE *fp = fopen (filename, "r");
 344   char *line = NULL;
 345   size_t line_len = 0;
 346   int modcounter = 0;
 347
 348   /* Don't complain if a file is not present or readable, simply silently
 349      ignore it.  */
 350   if (fp == NULL)
 351     return;
 352
 353   /* Process the known entries of the file.  Comments start with `#' and
 354      end with the end of the line.  Empty lines are ignored.  */
 355   while (!feof_unlocked (fp))
 356     {
 357       char *rp, *endp, *word;
 358       ssize_t n = __getdelim (&line, &line_len, '\n', fp);
 359       if (n < 0)
 360         /* An error occurred.  */
 361         break;
 362
 363       rp = line;
 364       /* Terminate the line (excluding comments or newline) by an NUL byte
 365          to simplify the following code.  */
 366       endp = strchr (rp, '#');
 367       if (endp != NULL)
 368         *endp = '\0';
 369       else
 370         if (rp[n - 1] == '\n')
 371           rp[n - 1] = '\0';
 372
 373       while (isspace (*rp))
 374         ++rp;
 375
 376       /* If this is an empty line go on with the next one.  */
 377       if (rp == endp)
 378         continue;
 379
 380       word = rp;
 381       while (*rp != '\0' && !isspace (*rp))
 382         ++rp;
 383
 384       if (rp - word == sizeof ("alias") - 1
 385           && memcmp (word, "alias", sizeof ("alias") - 1) == 0)
 386         add_alias (rp, *modules);
 387       else if (rp - word == sizeof ("module") - 1
 388                && memcmp (word, "module", sizeof ("module") - 1) == 0)
 389         add_module (rp, directory, dir_len, modules, nmodules, modcounter++);
 390       /* else */
 391         /* Otherwise ignore the line.  */
 392     }
 393
 394   free (line);
 395
 396   fclose (fp);
 397 }
 398
 399
 400 /* Determine the directories we are looking for data in.  */
 401 void
 402 __gconv_get_path (void)
 403 {
 404   struct path_elem *result;
 405   __libc_lock_define_initialized (static, lock);
 406
 407   __libc_lock_lock (lock);
 408
 409   /* Make sure there wasn't a second thread doing it already.  */
 410   result = (struct path_elem *) __gconv_path_elem;
 411   if (result == NULL)
 412     {
 413       /* Determine the complete path first.  */
 414       const char *user_path;
 415       char *gconv_path;
 416       size_t gconv_path_len;
 417       char *elem;
 418       char *oldp;
 419       char *cp;
 420       int nelems;
 421
 422       user_path = __secure_getenv ("GCONV_PATH");
 423       if (user_path == NULL)
 424         {
 425           /* No user-defined path.  Make a modifiable copy of the
 426              default path.  */
 427           gconv_path = strdupa (default_gconv_path);
 428           gconv_path_len = sizeof (default_gconv_path);
 429         }
 430       else
 431         {
 432           /* Append the default path to the user-defined path.  */
 433           size_t user_len = strlen (user_path);
 434
 435           gconv_path_len = user_len + 1 + sizeof (default_gconv_path);
 436           gconv_path = alloca (gconv_path_len);
 437           __mempcpy (__mempcpy (__mempcpy (gconv_path, user_path, user_len),
 438                                 ":", 1),
 439                      default_gconv_path, sizeof (default_gconv_path));
 440         }
 441
 442       /* In a first pass we calculate the number of elements.  */
 443       oldp = NULL;
 444       cp = strchr (gconv_path, ':');
 445       nelems = 1;
 446       while (cp != NULL)
 447         {
 448           if (cp != oldp + 1)
 449             ++nelems;
 450           oldp = cp;
 451           cp =  strchr (cp + 1, ':');
 452         }
 453
 454       /* Allocate the memory for the result.  */
 455       result = (struct path_elem *) malloc ((nelems + 1)
 456                                             * sizeof (struct path_elem)
 457                                             + gconv_path_len + nelems);
 458       if (result != NULL)
 459         {
 460           char *strspace = (char *) &result[nelems + 1];
 461           int n = 0;
 462
 463           /* Separate the individual parts.  */
 464           __gconv_max_path_elem_len = 0;
 465           elem = __strtok_r (gconv_path, ":", &gconv_path);
 466           assert (elem != NULL);
 467           do
 468             {
 469               result[n].name = strspace;
 470               strspace = __stpcpy (strspace, elem);
 471               if (strspace[-1] != '/')
 472                 *strspace++ = '/';
 473
 474               result[n].len = strspace - result[n].name;
 475               if (result[n].len > __gconv_max_path_elem_len)
 476                 __gconv_max_path_elem_len = result[n].len;
 477
 478               *strspace++ = '\0';
 479               ++n;
 480             }
 481           while ((elem = __strtok_r (NULL, ":", &gconv_path)) != NULL);
 482
 483           result[n].name = NULL;
 484           result[n].len = 0;
 485         }
 486
 487       __gconv_path_elem = result ?: &empty_path_elem;
 488     }
 489
 490   __libc_lock_unlock (lock);
 491 }
 492
 493
 494 /* Read all configuration files found in the user-specified and the default
 495    path.  */
 496 void
 497 __gconv_read_conf (void)
 498 {
 499   void *modules = NULL;
 500   size_t nmodules = 0;
 501   int save_errno = errno;
 502   size_t cnt;
 503
 504   /* Find out where we have to look.  */
 505   if (__gconv_path_elem == NULL)
 506     __gconv_get_path ();
 507
 508   for (cnt = 0; __gconv_path_elem[cnt].name != NULL; ++cnt)
 509     {
 510       char real_elem[__gconv_max_path_elem_len + sizeof (gconv_conf_filename)];
 511
 512       if (__realpath (__gconv_path_elem[cnt].name, real_elem) != NULL)
 513         {
 514           size_t elem_len = strlen (real_elem);
 515           char *filename;
 516
 517           filename = alloca (elem_len + 1 + sizeof (gconv_conf_filename));
 518           __mempcpy (__mempcpy (__mempcpy (filename, real_elem, elem_len),
 519                                 "/", 1),
 520                      gconv_conf_filename, sizeof (gconv_conf_filename));
 521
 522           /* Read the next configuration file.  */
 523           read_conf_file (filename, real_elem, elem_len, &modules, &nmodules);
 524         }
 525     }
 526
 527   /* Add the internal modules.  */
 528   for (cnt = 0; cnt < sizeof (builtin_modules) / sizeof (builtin_modules[0]);
 529        ++cnt)
 530     {
 531       struct gconv_alias fake_alias;
 532
 533       fake_alias.fromname = builtin_modules[cnt].from_string;
 534
 535       if (__tfind (&fake_alias, &__gconv_alias_db, __gconv_alias_compare)
 536           != NULL)
 537         /* It'll conflict so don't add it.  */
 538         continue;
 539
 540       insert_module (&builtin_modules[cnt]);
 541     }
 542
 543   /* Add aliases for builtin conversions.  */
 544   cnt = sizeof (builtin_aliases) / sizeof (builtin_aliases[0]);
 545   while (cnt > 0)
 546     {
 547       char *copy = strdupa (builtin_aliases[--cnt]);
 548       add_alias (copy, modules);
 549     }
 550
 551   /* Restore the error number.  */
 552   __set_errno (save_errno);
 553 }
 554
 555
 556
 557 /* Free all resources if necessary.  */
 558 static void __attribute__ ((unused))
 559 free_mem (void)
 560 {
 561   if (__gconv_path_elem != NULL && __gconv_path_elem != &empty_path_elem)
 562     free ((void *) __gconv_path_elem);
 563 }
 564
 565 text_set_element (__libc_subfreeres, free_mem);