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 <stddef.h>
  27 #include <stdio.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <unistd.h>
  31 #include <sys/param.h>
  32
  33 #include <bits/libc-lock.h>
  34 #include <gconv_int.h>
  35
  36
  37 /* This is the default path where we look for module lists.  */
  38 static const char default_gconv_path[] = GCONV_PATH;
  39
  40 /* The path elements, as determined by the __gconv_get_path function.
  41    All path elements end in a slash.  */
  42 struct path_elem *__gconv_path_elem;
  43 /* Maximum length of a single path element in __gconv_path_elem.  */
  44 size_t __gconv_max_path_elem_len;
  45
  46 /* We use the following struct if we couldn't allocate memory.  */
  47 static const struct path_elem empty_path_elem;
  48
  49 /* Name of the file containing the module information in the directories
  50    along the path.  */
  51 static const char gconv_conf_filename[] = "gconv-modules";
  52
  53 /* Filename extension for the modules.  */
  54 #ifndef MODULE_EXT
  55 # define MODULE_EXT ".so"
  56 #endif
  57 static const char gconv_module_ext[] = MODULE_EXT;
  58
  59 /* We have a few builtin transformations.  */
  60 static struct gconv_module builtin_modules[] =
  61 {
  62 #define BUILTIN_TRANSFORMATION(From, To, Cost, Name, Fct, Init, End, MinF, \
  63                                MaxF, MinT, MaxT) \
  64   {                                                                           \
  65     from_string: From,                                                        \
  66     to_string: To,                                                            \
  67     cost_hi: Cost,                                                            \
  68     cost_lo: INT_MAX,                                                         \
  69     module_name: Name                                                         \
  70   },
  71 #define BUILTIN_ALIAS(From, To)
  72
  73 #include "gconv_builtin.h"
  74 };
  75
  76 #undef BUILTIN_TRANSFORMATION
  77 #undef BUILTIN_ALIAS
  78
  79 static const char *builtin_aliases[] =
  80 {
  81 #define BUILTIN_TRANSFORMATION(From, To, Cost, Name, Fct, Init, End, MinF, \
  82                                MaxF, MinT, MaxT)
  83 #define BUILTIN_ALIAS(From, To) From " " To,
  84
  85 #include "gconv_builtin.h"
  86 };
  87
  88 #ifdef USE_IN_LIBIO
  89 # include <libio/libioP.h>
  90 # define __getdelim(line, len, c, fp) _IO_getdelim (line, len, c, fp)
  91 #endif
  92
  93
  94 /* Test whether there is already a matching module known.  */
  95 static int
  96 internal_function
  97 detect_conflict (const char *alias)
  98 {
  99   struct gconv_module *node = __gconv_modules_db;
 100
 101   while (node != NULL)
 102     {
 103       int cmpres = strcmp (alias, node->from_string);
 104
 105       if (cmpres == 0)
 106         /* We have a conflict.  */
 107         return 1;
 108       else if (cmpres < 0)
 109         node = node->left;
 110       else
 111         node = node->right;
 112     }
 113
 114   return node != NULL;
 115 }
 116
 117
 118 /* Add new alias.  */
 119 static inline void
 120 add_alias (char *rp, void *modules)
 121 {
 122   /* We now expect two more string.  The strings are normalized
 123      (converted to UPPER case) and strored in the alias database.  */
 124   struct gconv_alias *new_alias;
 125   char *from, *to, *wp;
 126
 127   while (isspace (*rp))
 128     ++rp;
 129   from = wp = rp;
 130   while (*rp != '\0' && !isspace (*rp))
 131     *wp++ = toupper (*rp++);
 132   if (*rp == '\0')
 133     /* There is no `to' string on the line.  Ignore it.  */
 134     return;
 135   *wp++ = '\0';
 136   to = ++rp;
 137   while (isspace (*rp))
 138     ++rp;
 139   while (*rp != '\0' && !isspace (*rp))
 140     *wp++ = toupper (*rp++);
 141   if (to == wp)
 142     /* No `to' string, ignore the line.  */
 143     return;
 144   *wp++ = '\0';
 145
 146   /* Test whether this alias conflicts with any available module.  */
 147   if (detect_conflict (from))
 148     /* It does conflict, don't add the alias.  */
 149     return;
 150
 151   new_alias = (struct gconv_alias *)
 152     malloc (sizeof (struct gconv_alias) + (wp - from));
 153   if (new_alias != NULL)
 154     {
 155       void **inserted;
 156
 157       new_alias->fromname = memcpy ((char *) new_alias
 158                                     + sizeof (struct gconv_alias),
 159                                     from, wp - from);
 160       new_alias->toname = new_alias->fromname + (to - from);
 161
 162       inserted = (void **) __tsearch (new_alias, &__gconv_alias_db,
 163                                       __gconv_alias_compare);
 164       if (inserted == NULL || *inserted != new_alias)
 165         /* Something went wrong, free this entry.  */
 166         free (new_alias);
 167     }
 168 }
 169
 170
 171 /* Insert a data structure for a new module in the search tree.  */
 172 static inline void
 173 internal_function
 174 insert_module (struct gconv_module *newp, int tobefreed)
 175 {
 176   struct gconv_module **rootp = &__gconv_modules_db;
 177
 178   while (*rootp != NULL)
 179     {
 180       struct gconv_module *root = *rootp;
 181       int cmpres;
 182
 183       cmpres = strcmp (newp->from_string, root->from_string);
 184       if (cmpres == 0)
 185         {
 186           /* Both strings are identical.  Insert the string at the
 187              end of the `same' list if it is not already there.  */
 188           while (strcmp (newp->from_string, root->from_string) != 0
 189                  || strcmp (newp->to_string, root->to_string) != 0)
 190             {
 191               rootp = &root->same;
 192               root = *rootp;
 193               if (root == NULL)
 194                 break;
 195             }
 196
 197           if (root != NULL)
 198             {
 199               /* This is a no new conversion.  But maybe the cost is
 200                  better.  */
 201               if (newp->cost_hi < root->cost_hi
 202                   || (newp->cost_hi == root->cost_hi
 203                       && newp->cost_lo < root->cost_lo))
 204                 {
 205                   root->cost_hi = newp->cost_hi;
 206                   root->cost_lo = newp->cost_lo;
 207                   root->module_name = newp->module_name;
 208                 }
 209
 210               if (tobefreed)
 211                 free (newp);
 212               return;
 213             }
 214
 215           break;
 216         }
 217       else if (cmpres < 0)
 218         rootp = &root->left;
 219       else
 220         rootp = &root->right;
 221     }
 222
 223   /* Plug in the new node here.  */
 224   *rootp = newp;
 225 }
 226
 227
 228 /* Add new module.  */
 229 static void
 230 internal_function
 231 add_module (char *rp, const char *directory, size_t dir_len, void **modules,
 232             size_t *nmodules, int modcounter)
 233 {
 234   /* We expect now
 235      1. `from' name
 236      2. `to' name
 237      3. filename of the module
 238      4. an optional cost value
 239   */
 240   struct gconv_alias fake_alias;
 241   struct gconv_module *new_module;
 242   char *from, *to, *module, *wp;
 243   int need_ext;
 244   int cost_hi;
 245
 246   while (isspace (*rp))
 247     ++rp;
 248   from = rp;
 249   while (*rp != '\0' && !isspace (*rp))
 250     {
 251       *rp = toupper (*rp);
 252       ++rp;
 253     }
 254   if (*rp == '\0')
 255     return;
 256   *rp++ = '\0';
 257   to = wp = rp;
 258   while (isspace (*rp))
 259     ++rp;
 260   while (*rp != '\0' && !isspace (*rp))
 261     *wp++ = toupper (*rp++);
 262   if (*rp == '\0')
 263     return;
 264   *wp++ = '\0';
 265   do
 266     ++rp;
 267   while (isspace (*rp));
 268   module = wp;
 269   while (*rp != '\0' && !isspace (*rp))
 270     *wp++ = *rp++;
 271   if (*rp == '\0')
 272     {
 273       /* There is no cost, use one by default.  */
 274       *wp++ = '\0';
 275       cost_hi = 1;
 276     }
 277   else
 278     {
 279       /* There might be a cost value.  */
 280       char *endp;
 281
 282       *wp++ = '\0';
 283       cost_hi = strtol (rp, &endp, 10);
 284       if (rp == endp || cost_hi < 1)
 285         /* No useful information.  */
 286         cost_hi = 1;
 287     }
 288
 289   if (module[0] == '\0')
 290     /* No module name given.  */
 291     return;
 292   if (module[0] == '/')
 293     dir_len = 0;
 294
 295   /* See whether we must add the ending.  */
 296   need_ext = 0;
 297   if (wp - module < (ptrdiff_t) sizeof (gconv_module_ext)
 298       || memcmp (wp - sizeof (gconv_module_ext), gconv_module_ext,
 299                  sizeof (gconv_module_ext)) != 0)
 300     /* We must add the module extension.  */
 301     need_ext = sizeof (gconv_module_ext) - 1;
 302
 303   /* See whether we have already an alias with this name defined.  */
 304   fake_alias.fromname = strndupa (from, to - from);
 305
 306   if (__tfind (&fake_alias, &__gconv_alias_db, __gconv_alias_compare) != NULL)
 307     /* This module duplicates an alias.  */
 308     return;
 309
 310   new_module = (struct gconv_module *) calloc (1,
 311                                                sizeof (struct gconv_module)
 312                                                + (wp - from)
 313                                                + dir_len + need_ext);
 314   if (new_module != NULL)
 315     {
 316       char *tmp;
 317
 318       new_module->from_string = tmp = (char *) (new_module + 1);
 319       tmp = __mempcpy (tmp, from, to - from);
 320
 321       new_module->to_string = tmp;
 322       tmp = __mempcpy (tmp, to, module - to);
 323
 324       new_module->cost_hi = cost_hi;
 325       new_module->cost_lo = modcounter;
 326
 327       new_module->module_name = tmp;
 328
 329       if (dir_len != 0)
 330         tmp = __mempcpy (tmp, directory, dir_len);
 331
 332       tmp = __mempcpy (tmp, module, wp - module);
 333
 334       if (need_ext)
 335         memcpy (tmp - 1, gconv_module_ext, sizeof (gconv_module_ext));
 336
 337       /* Now insert the new module data structure in our search tree.  */
 338       insert_module (new_module, 1);
 339     }
 340 }
 341
 342
 343 /* Read the next configuration file.  */
 344 static void
 345 internal_function
 346 read_conf_file (const char *filename, const char *directory, size_t dir_len,
 347                 void **modules, size_t *nmodules)
 348 {
 349   FILE *fp = fopen (filename, "r");
 350   char *line = NULL;
 351   size_t line_len = 0;
 352   int modcounter = 0;
 353
 354   /* Don't complain if a file is not present or readable, simply silently
 355      ignore it.  */
 356   if (fp == NULL)
 357     return;
 358
 359   /* Process the known entries of the file.  Comments start with `#' and
 360      end with the end of the line.  Empty lines are ignored.  */
 361   while (!feof_unlocked (fp))
 362     {
 363       char *rp, *endp, *word;
 364       ssize_t n = __getdelim (&line, &line_len, '\n', fp);
 365       if (n < 0)
 366         /* An error occurred.  */
 367         break;
 368
 369       rp = line;
 370       /* Terminate the line (excluding comments or newline) by an NUL byte
 371          to simplify the following code.  */
 372       endp = strchr (rp, '#');
 373       if (endp != NULL)
 374         *endp = '\0';
 375       else
 376         if (rp[n - 1] == '\n')
 377           rp[n - 1] = '\0';
 378
 379       while (isspace (*rp))
 380         ++rp;
 381
 382       /* If this is an empty line go on with the next one.  */
 383       if (rp == endp)
 384         continue;
 385
 386       word = rp;
 387       while (*rp != '\0' && !isspace (*rp))
 388         ++rp;
 389
 390       if (rp - word == sizeof ("alias") - 1
 391           && memcmp (word, "alias", sizeof ("alias") - 1) == 0)
 392         add_alias (rp, *modules);
 393       else if (rp - word == sizeof ("module") - 1
 394                && memcmp (word, "module", sizeof ("module") - 1) == 0)
 395         add_module (rp, directory, dir_len, modules, nmodules, modcounter++);
 396       /* else */
 397         /* Otherwise ignore the line.  */
 398     }
 399
 400   free (line);
 401
 402   fclose (fp);
 403 }
 404
 405
 406 /* Determine the directories we are looking for data in.  */
 407 void
 408 __gconv_get_path (void)
 409 {
 410   struct path_elem *result;
 411   __libc_lock_define_initialized (static, lock);
 412
 413   __libc_lock_lock (lock);
 414
 415   /* Make sure there wasn't a second thread doing it already.  */
 416   result = (struct path_elem *) __gconv_path_elem;
 417   if (result == NULL)
 418     {
 419       /* Determine the complete path first.  */
 420       const char *user_path;
 421       char *gconv_path;
 422       size_t gconv_path_len;
 423       char *elem;
 424       char *oldp;
 425       char *cp;
 426       int nelems;
 427       char *cwd;
 428       size_t cwdlen;
 429
 430       user_path = getenv ("GCONV_PATH");
 431       if (user_path == NULL)
 432         {
 433           /* No user-defined path.  Make a modifiable copy of the
 434              default path.  */
 435           gconv_path = strdupa (default_gconv_path);
 436           gconv_path_len = sizeof (default_gconv_path);
 437           cwd = NULL;
 438           cwdlen = 0;
 439         }
 440       else
 441         {
 442           /* Append the default path to the user-defined path.  */
 443           size_t user_len = strlen (user_path);
 444
 445           gconv_path_len = user_len + 1 + sizeof (default_gconv_path);
 446           gconv_path = alloca (gconv_path_len);
 447           __mempcpy (__mempcpy (__mempcpy (gconv_path, user_path, user_len),
 448                                 ":", 1),
 449                      default_gconv_path, sizeof (default_gconv_path));
 450           cwd = __getcwd (NULL, 0);
 451           cwdlen = strlen (cwd);
 452         }
 453       assert (default_gconv_path[0] == '/');
 454
 455       /* In a first pass we calculate the number of elements.  */
 456       oldp = NULL;
 457       cp = strchr (gconv_path, ':');
 458       nelems = 1;
 459       while (cp != NULL)
 460         {
 461           if (cp != oldp + 1)
 462             ++nelems;
 463           oldp = cp;
 464           cp =  strchr (cp + 1, ':');
 465         }
 466
 467       /* Allocate the memory for the result.  */
 468       result = (struct path_elem *) malloc ((nelems + 1)
 469                                             * sizeof (struct path_elem)
 470                                             + gconv_path_len + nelems
 471                                             + (nelems - 1) * (cwdlen + 1));
 472       if (result != NULL)
 473         {
 474           char *strspace = (char *) &result[nelems + 1];
 475           int n = 0;
 476
 477           /* Separate the individual parts.  */
 478           __gconv_max_path_elem_len = 0;
 479           elem = __strtok_r (gconv_path, ":", &gconv_path);
 480           assert (elem != NULL);
 481           do
 482             {
 483               result[n].name = strspace;
 484               if (elem[0] != '/')
 485                 {
 486                   assert (cwd != NULL);
 487                   strspace = __mempcpy (strspace, cwd, cwdlen);
 488                   *strspace++ = '/';
 489                 }
 490               strspace = __stpcpy (strspace, elem);
 491               if (strspace[-1] != '/')
 492                 *strspace++ = '/';
 493
 494               result[n].len = strspace - result[n].name;
 495               if (result[n].len > __gconv_max_path_elem_len)
 496                 __gconv_max_path_elem_len = result[n].len;
 497
 498               *strspace++ = '\0';
 499               ++n;
 500             }
 501           while ((elem = __strtok_r (NULL, ":", &gconv_path)) != NULL);
 502
 503           result[n].name = NULL;
 504           result[n].len = 0;
 505         }
 506
 507       __gconv_path_elem = result ?: &empty_path_elem;
 508
 509       if (cwd != NULL)
 510         free (cwd);
 511     }
 512
 513   __libc_lock_unlock (lock);
 514 }
 515
 516
 517 /* Read all configuration files found in the user-specified and the default
 518    path.  */
 519 void
 520 __gconv_read_conf (void)
 521 {
 522   void *modules = NULL;
 523   size_t nmodules = 0;
 524   int save_errno = errno;
 525   size_t cnt;
 526
 527   /* Find out where we have to look.  */
 528   if (__gconv_path_elem == NULL)
 529     __gconv_get_path ();
 530
 531   for (cnt = 0; __gconv_path_elem[cnt].name != NULL; ++cnt)
 532     {
 533       const char *elem = __gconv_path_elem[cnt].name;
 534       size_t elem_len = __gconv_path_elem[cnt].len;
 535       char *filename;
 536
 537       /* No slash needs to be inserted between elem and gconv_conf_filename;
 538          elem already ends in a slash.  */
 539       filename = alloca (elem_len + sizeof (gconv_conf_filename));
 540       __mempcpy (__mempcpy (filename, elem, elem_len),
 541                  gconv_conf_filename, sizeof (gconv_conf_filename));
 542
 543       /* Read the next configuration file.  */
 544       read_conf_file (filename, elem, elem_len, &modules, &nmodules);
 545     }
 546
 547   /* Add the internal modules.  */
 548   for (cnt = 0; cnt < sizeof (builtin_modules) / sizeof (builtin_modules[0]);
 549        ++cnt)
 550     {
 551       struct gconv_alias fake_alias;
 552
 553       fake_alias.fromname = builtin_modules[cnt].from_string;
 554
 555       if (__tfind (&fake_alias, &__gconv_alias_db, __gconv_alias_compare)
 556           != NULL)
 557         /* It'll conflict so don't add it.  */
 558         continue;
 559
 560       insert_module (&builtin_modules[cnt], 0);
 561     }
 562
 563   /* Add aliases for builtin conversions.  */
 564   cnt = sizeof (builtin_aliases) / sizeof (builtin_aliases[0]);
 565   while (cnt > 0)
 566     {
 567       char *copy = strdupa (builtin_aliases[--cnt]);
 568       add_alias (copy, modules);
 569     }
 570
 571   /* Restore the error number.  */
 572   __set_errno (save_errno);
 573 }
 574
 575
 576
 577 /* Free all resources if necessary.  */
 578 static void __attribute__ ((unused))
 579 free_mem (void)
 580 {
 581   if (__gconv_path_elem != NULL && __gconv_path_elem != &empty_path_elem)
 582     free ((void *) __gconv_path_elem);
 583 }
 584
 585 text_set_element (__libc_subfreeres, free_mem);