libcpp/makeucnid.c

   1 /* Make ucnid.h from various sources.
   2    Copyright (C) 2005 Free Software Foundation, Inc.
   3
   4 This program is free software; you can redistribute it and/or modify it
   5 under the terms of the GNU General Public License as published by the
   6 Free Software Foundation; either version 2, or (at your option) any
   7 later version.
   8
   9 This program 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
  12 GNU General Public License for more details.
  13
  14 You should have received a copy of the GNU General Public License
  15 along with this program; if not, write to the Free Software
  16 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 /* Run this program as
  19    ./makeucnid ucnid.tab UnicodeData.txt DerivedNormalizationProps.txt \
  20        > ucnid.h
  21 */
  22
  23 #include <stdio.h>
  24 #include <string.h>
  25 #include <ctype.h>
  26 #include <stdbool.h>
  27 #include <stdlib.h>
  28
  29 enum {
  30   C99 = 1,
  31   CXX = 2,
  32   digit = 4,
  33   not_NFC = 8,
  34   not_NFKC = 16,
  35   maybe_not_NFC = 32
  36 };
  37
  38 static unsigned flags[65536];
  39 static unsigned short decomp[65536][2];
  40 static unsigned char combining_value[65536];
  41
  42 /* Die!  */
  43
  44 static void
  45 fail (const char *s)
  46 {
  47   fprintf (stderr, "%s\n", s);
  48   exit (1);
  49 }
  50
  51 /* Read ucnid.tab and set the C99 and CXX flags in header[].  */
  52
  53 static void
  54 read_ucnid (const char *fname)
  55 {
  56   FILE *f = fopen (fname, "r");
  57   unsigned fl = 0;
  58
  59   if (!f)
  60     fail ("opening ucnid.tab");
  61   for (;;)
  62     {
  63       char line[256];
  64
  65       if (!fgets (line, sizeof (line), f))
  66         break;
  67       if (strcmp (line, "[C99]\n") == 0)
  68         fl = C99;
  69       else if (strcmp (line, "[CXX]\n") == 0)
  70         fl = CXX;
  71       else if (isxdigit (line[0]))
  72         {
  73           char *l = line;
  74           while (*l)
  75             {
  76               unsigned long start, end;
  77               char *endptr;
  78               start = strtoul (l, &endptr, 16);
  79               if (endptr == l || (*endptr != '-' && ! isspace (*endptr)))
  80                 fail ("parsing ucnid.tab [1]");
  81               l = endptr;
  82               if (*l != '-')
  83                 end = start;
  84               else
  85                 {
  86                   end = strtoul (l + 1, &endptr, 16);
  87                   if (end < start)
  88                     fail ("parsing ucnid.tab, end before start");
  89                   l = endptr;
  90                   if (! isspace (*l))
  91                     fail ("parsing ucnid.tab, junk after range");
  92                 }
  93               while (isspace (*l))
  94                 l++;
  95               if (end > 0xFFFF)
  96                 fail ("parsing ucnid.tab, end too large");
  97               while (start <= end)
  98                 flags[start++] |= fl;
  99             }
 100         }
 101     }
 102   if (ferror (f))
 103     fail ("reading ucnid.tab");
 104   fclose (f);
 105 }
 106
 107 /* Read UnicodeData.txt and set the 'digit' flag, and
 108    also fill in the 'decomp' table to be the decompositions of
 109    characters for which both the character decomposed and all the code
 110    points in the decomposition are either C99 or CXX.  */
 111
 112 static void
 113 read_table (char *fname)
 114 {
 115   FILE * f = fopen (fname, "r");
 116
 117   if (!f)
 118     fail ("opening UnicodeData.txt");
 119   for (;;)
 120     {
 121       char line[256];
 122       unsigned long codepoint, this_decomp[4];
 123       char *l;
 124       int i;
 125       int decomp_useful;
 126
 127       if (!fgets (line, sizeof (line), f))
 128         break;
 129       codepoint = strtoul (line, &l, 16);
 130       if (l == line || *l != ';')
 131         fail ("parsing UnicodeData.txt, reading code point");
 132       if (codepoint > 0xffff || ! (flags[codepoint] & (C99 | CXX)))
 133         continue;
 134
 135       do {
 136         l++;
 137       } while (*l != ';');
 138       /* Category value; things starting with 'N' are numbers of some
 139          kind.  */
 140       if (*++l == 'N')
 141         flags[codepoint] |= digit;
 142
 143       do {
 144         l++;
 145       } while (*l != ';');
 146       /* Canonical combining class; in NFC/NFKC, they must be increasing
 147          (or zero).  */
 148       if (! isdigit (*++l))
 149         fail ("parsing UnicodeData.txt, combining class not number");
 150       combining_value[codepoint] = strtoul (l, &l, 10);
 151       if (*l++ != ';')
 152         fail ("parsing UnicodeData.txt, junk after combining class");
 153
 154       /* Skip over bidi value.  */
 155       do {
 156         l++;
 157       } while (*l != ';');
 158
 159       /* Decomposition mapping.  */
 160       decomp_useful = flags[codepoint];
 161       if (*++l == '<')  /* Compatibility mapping. */
 162         continue;
 163       for (i = 0; i < 4; i++)
 164         {
 165           if (*l == ';')
 166             break;
 167           if (!isxdigit (*l))
 168             fail ("parsing UnicodeData.txt, decomposition format");
 169           this_decomp[i] = strtoul (l, &l, 16);
 170           decomp_useful &= flags[this_decomp[i]];
 171           while (isspace (*l))
 172             l++;
 173         }
 174       if (i > 2)  /* Decomposition too long.  */
 175         fail ("parsing UnicodeData.txt, decomposition too long");
 176       if (decomp_useful)
 177         while (--i >= 0)
 178           decomp[codepoint][i] = this_decomp[i];
 179     }
 180   if (ferror (f))
 181     fail ("reading UnicodeData.txt");
 182   fclose (f);
 183 }
 184
 185 /* Read DerivedNormalizationProps.txt and set the flags that say whether
 186    a character is in NFC, NFKC, or is context-dependent.  */
 187
 188 static void
 189 read_derived (const char *fname)
 190 {
 191   FILE * f = fopen (fname, "r");
 192
 193   if (!f)
 194     fail ("opening DerivedNormalizationProps.txt");
 195   for (;;)
 196     {
 197       char line[256];
 198       unsigned long start, end;
 199       char *l;
 200       bool not_NFC_p, not_NFKC_p, maybe_not_NFC_p;
 201
 202       if (!fgets (line, sizeof (line), f))
 203         break;
 204       not_NFC_p = (strstr (line, "; NFC_QC; N") != NULL);
 205       not_NFKC_p = (strstr (line, "; NFKC_QC; N") != NULL);
 206       maybe_not_NFC_p = (strstr (line, "; NFC_QC; M") != NULL);
 207       if (! not_NFC_p && ! not_NFKC_p && ! maybe_not_NFC_p)
 208         continue;
 209
 210       start = strtoul (line, &l, 16);
 211       if (l == line)
 212         fail ("parsing DerivedNormalizationProps.txt, reading start");
 213       if (start > 0xffff)
 214         continue;
 215       if (*l == '.' && l[1] == '.')
 216         end = strtoul (l + 2, &l, 16);
 217       else
 218         end = start;
 219
 220       while (start <= end)
 221         flags[start++] |= ((not_NFC_p ? not_NFC : 0)
 222                            | (not_NFKC_p ? not_NFKC : 0)
 223                            | (maybe_not_NFC_p ? maybe_not_NFC : 0)
 224                            );
 225     }
 226   if (ferror (f))
 227     fail ("reading DerivedNormalizationProps.txt");
 228   fclose (f);
 229 }
 230
 231 /* Write out the table.
 232    The table consists of two words per entry.  The first word is the flags
 233    for the unicode code points up to and including the second word.  */
 234
 235 static void
 236 write_table (void)
 237 {
 238   unsigned i;
 239   unsigned last_flag = flags[0];
 240   bool really_safe = decomp[0][0] == 0;
 241   unsigned char last_combine = combining_value[0];
 242
 243   for (i = 1; i <= 65536; i++)
 244     if (i == 65536
 245         || (flags[i] != last_flag && ((flags[i] | last_flag) & (C99 | CXX)))
 246         || really_safe != (decomp[i][0] == 0)
 247         || combining_value[i] != last_combine)
 248       {
 249         printf ("{ %s|%s|%s|%s|%s|%s|%s, %3d, %#06x },\n",
 250                 last_flag & C99 ? "C99" : "  0",
 251                 last_flag & digit ? "DIG" : "  0",
 252                 last_flag & CXX ? "CXX" : "  0",
 253                 really_safe ? "CID" : "  0",
 254                 last_flag & not_NFC ? "  0" : "NFC",
 255                 last_flag & not_NFKC ? "  0" : "NKC",
 256                 last_flag & maybe_not_NFC ? "CTX" : "  0",
 257                 combining_value[i - 1],
 258                 i - 1);
 259         last_flag = flags[i];
 260         last_combine = combining_value[0];
 261         really_safe = decomp[i][0] == 0;
 262       }
 263 }
 264
 265 /* Print out the huge copyright notice.  */
 266
 267 static void
 268 write_copyright (void)
 269 {
 270   static const char copyright[] = "\
 271 /* Unicode characters and various properties.\n\
 272    Copyright (C) 2003, 2005 Free Software Foundation, Inc.\n\
 273 \n\
 274    This program is free software; you can redistribute it and/or modify it\n\
 275    under the terms of the GNU General Public License as published by the\n\
 276    Free Software Foundation; either version 2, or (at your option) any\n\
 277    later version.\n\
 278 \n\
 279    This program is distributed in the hope that it will be useful,\n\
 280    but WITHOUT ANY WARRANTY; without even the implied warranty of\n\
 281    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n\
 282    GNU General Public License for more details.\n\
 283 \n\
 284    You should have received a copy of the GNU General Public License\n\
 285    along with this program; if not, write to the Free Software\n\
 286    Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.\n\
 287 \n\
 288 \n\
 289    Copyright (C) 1991-2005 Unicode, Inc.  All rights reserved.\n\
 290    Distributed under the Terms of Use in\n\
 291    http://www.unicode.org/copyright.html.\n\
 292 \n\
 293    Permission is hereby granted, free of charge, to any person\n\
 294    obtaining a copy of the Unicode data files and any associated\n\
 295    documentation (the \"Data Files\") or Unicode software and any\n\
 296    associated documentation (the \"Software\") to deal in the Data Files\n\
 297    or Software without restriction, including without limitation the\n\
 298    rights to use, copy, modify, merge, publish, distribute, and/or\n\
 299    sell copies of the Data Files or Software, and to permit persons to\n\
 300    whom the Data Files or Software are furnished to do so, provided\n\
 301    that (a) the above copyright notice(s) and this permission notice\n\
 302    appear with all copies of the Data Files or Software, (b) both the\n\
 303    above copyright notice(s) and this permission notice appear in\n\
 304    associated documentation, and (c) there is clear notice in each\n\
 305    modified Data File or in the Software as well as in the\n\
 306    documentation associated with the Data File(s) or Software that the\n\
 307    data or software has been modified.\n\
 308 \n\
 309    THE DATA FILES AND SOFTWARE ARE PROVIDED \"AS IS\", WITHOUT WARRANTY\n\
 310    OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE\n\
 311    WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND\n\
 312    NONINFRINGEMENT OF THIRD PARTY RIGHTS. IN NO EVENT SHALL THE\n\
 313    COPYRIGHT HOLDER OR HOLDERS INCLUDED IN THIS NOTICE BE LIABLE FOR\n\
 314    ANY CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY\n\
 315    DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,\n\
 316    WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS\n\
 317    ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE\n\
 318    OF THE DATA FILES OR SOFTWARE.\n\
 319 \n\
 320    Except as contained in this notice, the name of a copyright holder\n\
 321    shall not be used in advertising or otherwise to promote the sale,\n\
 322    use or other dealings in these Data Files or Software without prior\n\
 323    written authorization of the copyright holder.  */\n";
 324
 325    puts (copyright);
 326 }
 327
 328 /* Main program.  */
 329
 330 int
 331 main(int argc, char ** argv)
 332 {
 333   if (argc != 4)
 334     fail ("too few arguments to makeucn");
 335   read_ucnid (argv[1]);
 336   read_table (argv[2]);
 337   read_derived (argv[3]);
 338
 339   write_copyright ();
 340   write_table ();
 341   return 0;
 342 }