localedata/tst-ctype.c

   1 /* Copyright (C) 2000-2014 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Contributed by Ulrich Drepper <drepper@gnu.org>, 2000.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <http://www.gnu.org/licenses/>.  */
  18
  19 #include <ctype.h>
  20 #include <locale.h>
  21 #include <langinfo.h>
  22 #include <stdio.h>
  23 #include <string.h>
  24
  25
  26 static const char lower[] = "abcdefghijklmnopqrstuvwxyz";
  27 static const char upper[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
  28 static const char digits[] = "0123456789";
  29 static const char cntrl[] = "\
  30 \x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f\
  31 \x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f ";
  32
  33
  34 static struct classes
  35 {
  36   const char *name;
  37   int mask;
  38 } classes[] =
  39 {
  40 #define ENTRY(name) { #name, _IS##name }
  41   ENTRY (upper),
  42   ENTRY (lower),
  43   ENTRY (alpha),
  44   ENTRY (digit),
  45   ENTRY (xdigit),
  46   ENTRY (space),
  47   ENTRY (print),
  48   ENTRY (graph),
  49   ENTRY (blank),
  50   ENTRY (cntrl),
  51   ENTRY (punct),
  52   ENTRY (alnum)
  53 };
  54 #define nclasses (sizeof (classes) / sizeof (classes[0]))
  55
  56
  57 #define FAIL(str, args...) \
  58   {                                                                           \
  59     printf ("      " str "\n", ##args);                                       \
  60     ++errors;                                                                 \
  61   }
  62
  63
  64 int
  65 main (void)
  66 {
  67   const char *cp;
  68   const char *cp2;
  69   int errors = 0;
  70   char *inpline = NULL;
  71   size_t inplinelen = 0;
  72   char *resline = NULL;
  73   size_t reslinelen = 0;
  74   size_t n;
  75   const unsigned short int *__ctype_b;
  76
  77   setlocale (LC_ALL, "");
  78
  79   printf ("Testing the ctype data of the `%s' locale\n",
  80           setlocale (LC_CTYPE, NULL));
  81
  82   __ctype_b = ((const unsigned short *) nl_langinfo (_NL_CTYPE_CLASS)) + 128;
  83
  84 #if 0
  85   /* Just for debugging.  */
  86
  87   /* Contents of the class array.  */
  88   printf ("\
  89 upper = %04x  lower = %04x  alpha = %04x  digit = %04x  xdigit = %04x\n\
  90 space = %04x  print = %04x  graph = %04x  blank = %04x  cntrl  = %04x\n\
  91 punct = %04x  alnum = %04x\n",
  92           _ISupper, _ISlower, _ISalpha, _ISdigit, _ISxdigit,
  93           _ISspace, _ISprint, _ISgraph, _ISblank, _IScntrl,
  94           _ISpunct, _ISalnum);
  95
  96   while (n < 256)
  97     {
  98       if (n % 8 == 0)
  99         printf ("%02x: ", n);
 100       printf ("%04x%s", __ctype_b[n], (n + 1) % 8 == 0 ? "\n" : " ");
 101       ++n;
 102     }
 103 #endif
 104
 105   puts ("  Test of ASCII character range\n    special NUL byte handling");
 106   if (isupper ('\0'))
 107     FAIL ("isupper ('\\0') is true");
 108   if (islower ('\0'))
 109     FAIL ("islower ('\\0') is true");
 110   if (isalpha ('\0'))
 111     FAIL ("isalpha ('\\0') is true");
 112   if (isdigit ('\0'))
 113     FAIL ("isdigit ('\\0') is true");
 114   if (isxdigit ('\0'))
 115     FAIL ("isxdigit ('\\0') is true");
 116   if (isspace ('\0'))
 117     FAIL ("isspace ('\\0') is true");
 118   if (isprint ('\0'))
 119     FAIL ("isprint ('\\0') is true");
 120   if (isgraph ('\0'))
 121     FAIL ("isgraph ('\\0') is true");
 122   if (isblank ('\0'))
 123     FAIL ("isblank ('\\0') is true");
 124   if (! iscntrl ('\0'))
 125     FAIL ("iscntrl ('\\0') not true");
 126   if (ispunct ('\0'))
 127     FAIL ("ispunct ('\\0') is true");
 128   if (isalnum ('\0'))
 129     FAIL ("isalnum ('\\0') is true");
 130
 131   puts ("    islower()");
 132   for (cp = lower; *cp != '\0'; ++cp)
 133     if (! islower (*cp))
 134       FAIL ("islower ('%c') not true", *cp);
 135   for (cp = upper; *cp != '\0'; ++cp)
 136     if (islower (*cp))
 137       FAIL ("islower ('%c') is true", *cp);
 138   for (cp = digits; *cp != '\0'; ++cp)
 139     if (islower (*cp))
 140       FAIL ("islower ('%c') is true", *cp);
 141   for (cp = cntrl; *cp != '\0'; ++cp)
 142     if (islower (*cp))
 143       FAIL ("islower ('\\x%02x') is true", *cp);
 144
 145   puts ("    isupper()");
 146   for (cp = lower; *cp != '\0'; ++cp)
 147     if (isupper (*cp))
 148       FAIL ("isupper ('%c') is true", *cp);
 149   for (cp = upper; *cp != '\0'; ++cp)
 150     if (! isupper (*cp))
 151       FAIL ("isupper ('%c') not true", *cp);
 152   for (cp = digits; *cp != '\0'; ++cp)
 153     if (isupper (*cp))
 154       FAIL ("isupper ('%c') is true", *cp);
 155   for (cp = cntrl; *cp != '\0'; ++cp)
 156     if (isupper (*cp))
 157       FAIL ("isupper ('\\x%02x') is true", *cp);
 158
 159   puts ("    isalpha()");
 160   for (cp = lower; *cp != '\0'; ++cp)
 161     if (! isalpha (*cp))
 162       FAIL ("isalpha ('%c') not true", *cp);
 163   for (cp = upper; *cp != '\0'; ++cp)
 164     if (! isalpha (*cp))
 165       FAIL ("isalpha ('%c') not true", *cp);
 166   for (cp = digits; *cp != '\0'; ++cp)
 167     if (isalpha (*cp))
 168       FAIL ("isalpha ('%c') is true", *cp);
 169   for (cp = cntrl; *cp != '\0'; ++cp)
 170     if (isalpha (*cp))
 171       FAIL ("isalpha ('\\x%02x') is true", *cp);
 172
 173   puts ("    isdigit()");
 174   for (cp = lower; *cp != '\0'; ++cp)
 175     if (isdigit (*cp))
 176       FAIL ("isdigit ('%c') is true", *cp);
 177   for (cp = upper; *cp != '\0'; ++cp)
 178     if (isdigit (*cp))
 179       FAIL ("isdigit ('%c') is true", *cp);
 180   for (cp = digits; *cp != '\0'; ++cp)
 181     if (! isdigit (*cp))
 182       FAIL ("isdigit ('%c') not true", *cp);
 183   for (cp = cntrl; *cp != '\0'; ++cp)
 184     if (isdigit (*cp))
 185       FAIL ("isdigit ('\\x%02x') is true", *cp);
 186
 187   puts ("    isxdigit()");
 188   for (cp = lower; *cp != '\0'; ++cp)
 189     if ((! isxdigit (*cp) && cp - lower < 6)
 190         || (isxdigit (*cp) && cp - lower >= 6))
 191       FAIL ("isxdigit ('%c') %s true", *cp, cp - upper < 6 ? "not" : "is");
 192   for (cp = upper; *cp != '\0'; ++cp)
 193     if ((! isxdigit (*cp) && cp - upper < 6)
 194         || (isxdigit (*cp) && cp - upper >= 6))
 195       FAIL ("isxdigit ('%c') %s true", *cp, cp - upper < 6 ? "not" : "is");
 196   for (cp = digits; *cp != '\0'; ++cp)
 197     if (! isxdigit (*cp))
 198       FAIL ("isxdigit ('%c') not true", *cp);
 199   for (cp = cntrl; *cp != '\0'; ++cp)
 200     if (isxdigit (*cp))
 201       FAIL ("isxdigit ('\\x%02x') is true", *cp);
 202
 203   puts ("    isspace()");
 204   for (cp = lower; *cp != '\0'; ++cp)
 205     if (isspace (*cp))
 206       FAIL ("isspace ('%c') is true", *cp);
 207   for (cp = upper; *cp != '\0'; ++cp)
 208     if (isspace (*cp))
 209       FAIL ("isspace ('%c') is true", *cp);
 210   for (cp = digits; *cp != '\0'; ++cp)
 211     if (isspace (*cp))
 212       FAIL ("isspace ('%c') is true", *cp);
 213   for (cp = cntrl; *cp != '\0'; ++cp)
 214     if ((isspace (*cp) && ((*cp < '\x09' || *cp > '\x0d') && *cp != ' '))
 215         || (! isspace (*cp)
 216             && ((*cp >= '\x09' && *cp <= '\x0d') || *cp == ' ')))
 217       FAIL ("isspace ('\\x%02x') %s true", *cp,
 218             (*cp < '\x09' || *cp > '\x0d') ? "is" : "not");
 219
 220   puts ("    isprint()");
 221   for (cp = lower; *cp != '\0'; ++cp)
 222     if (! isprint (*cp))
 223       FAIL ("isprint ('%c') not true", *cp);
 224   for (cp = upper; *cp != '\0'; ++cp)
 225     if (! isprint (*cp))
 226       FAIL ("isprint ('%c') not true", *cp);
 227   for (cp = digits; *cp != '\0'; ++cp)
 228     if (! isprint (*cp))
 229       FAIL ("isprint ('%c') not true", *cp);
 230   for (cp = cntrl; *cp != '\0'; ++cp)
 231     if ((isprint (*cp) && *cp != ' ')
 232         || (! isprint (*cp) && *cp == ' '))
 233       FAIL ("isprint ('\\x%02x') is true", *cp);
 234
 235   puts ("    isgraph()");
 236   for (cp = lower; *cp != '\0'; ++cp)
 237     if (! isgraph (*cp))
 238       FAIL ("isgraph ('%c') not true", *cp);
 239   for (cp = upper; *cp != '\0'; ++cp)
 240     if (! isgraph (*cp))
 241       FAIL ("isgraph ('%c') not true", *cp);
 242   for (cp = digits; *cp != '\0'; ++cp)
 243     if (! isgraph (*cp))
 244       FAIL ("isgraph ('%c') not true", *cp);
 245   for (cp = cntrl; *cp != '\0'; ++cp)
 246     if (isgraph (*cp))
 247       FAIL ("isgraph ('\\x%02x') is true", *cp);
 248
 249   puts ("    isblank()");
 250   for (cp = lower; *cp != '\0'; ++cp)
 251     if (isblank (*cp))
 252       FAIL ("isblank ('%c') is true", *cp);
 253   for (cp = upper; *cp != '\0'; ++cp)
 254     if (isblank (*cp))
 255       FAIL ("isblank ('%c') is true", *cp);
 256   for (cp = digits; *cp != '\0'; ++cp)
 257     if (isblank (*cp))
 258       FAIL ("isblank ('%c') is true", *cp);
 259   for (cp = cntrl; *cp != '\0'; ++cp)
 260     if ((isblank (*cp) && *cp != '\x09' && *cp != ' ')
 261         || (! isblank (*cp) && (*cp == '\x09' || *cp == ' ')))
 262       FAIL ("isblank ('\\x%02x') %s true", *cp, *cp != '\x09' ? "is" : "not");
 263
 264   puts ("    iscntrl()");
 265   for (cp = lower; *cp != '\0'; ++cp)
 266     if (iscntrl (*cp))
 267       FAIL ("iscntrl ('%c') is true", *cp);
 268   for (cp = upper; *cp != '\0'; ++cp)
 269     if (iscntrl (*cp))
 270       FAIL ("iscntrl ('%c') is true", *cp);
 271   for (cp = digits; *cp != '\0'; ++cp)
 272     if (iscntrl (*cp))
 273       FAIL ("iscntrl ('%c') is true", *cp);
 274   for (cp = cntrl; *cp != '\0'; ++cp)
 275     if ((iscntrl (*cp) && *cp == ' ')
 276         || (! iscntrl (*cp) && *cp != ' '))
 277       FAIL ("iscntrl ('\\x%02x') not true", *cp);
 278
 279   puts ("    ispunct()");
 280   for (cp = lower; *cp != '\0'; ++cp)
 281     if (ispunct (*cp))
 282       FAIL ("ispunct ('%c') is true", *cp);
 283   for (cp = upper; *cp != '\0'; ++cp)
 284     if (ispunct (*cp))
 285       FAIL ("ispunct ('%c') is true", *cp);
 286   for (cp = digits; *cp != '\0'; ++cp)
 287     if (ispunct (*cp))
 288       FAIL ("ispunct ('%c') is true", *cp);
 289   for (cp = cntrl; *cp != '\0'; ++cp)
 290     if (ispunct (*cp))
 291       FAIL ("ispunct ('\\x%02x') is true", *cp);
 292
 293   puts ("    isalnum()");
 294   for (cp = lower; *cp != '\0'; ++cp)
 295     if (! isalnum (*cp))
 296       FAIL ("isalnum ('%c') not true", *cp);
 297   for (cp = upper; *cp != '\0'; ++cp)
 298     if (! isalnum (*cp))
 299       FAIL ("isalnum ('%c') not true", *cp);
 300   for (cp = digits; *cp != '\0'; ++cp)
 301     if (! isalnum (*cp))
 302       FAIL ("isalnum ('%c') not true", *cp);
 303   for (cp = cntrl; *cp != '\0'; ++cp)
 304     if (isalnum (*cp))
 305       FAIL ("isalnum ('\\x%02x') is true", *cp);
 306
 307
 308   puts ("    tolower()");
 309   for (cp = lower; *cp != '\0'; ++cp)
 310     if (tolower (*cp) != *cp)
 311       FAIL ("tolower ('%c') != '%c'", *cp, *cp);
 312   for (cp = upper, cp2 = lower; *cp != '\0'; ++cp, ++cp2)
 313     if (tolower (*cp) != *cp2)
 314       FAIL ("tolower ('%c') != '%c'", *cp, *cp2);
 315   for (cp = digits; *cp != '\0'; ++cp)
 316     if (tolower (*cp) != *cp)
 317       FAIL ("tolower ('%c') != '%c'", *cp, *cp);
 318   for (cp = cntrl; *cp != '\0'; ++cp)
 319     if (tolower (*cp) != *cp)
 320       FAIL ("tolower ('\\x%02x') != '\\x%02x'", *cp, *cp);
 321
 322   puts ("    toupper()");
 323   for (cp = lower, cp2 = upper; *cp != '\0'; ++cp, ++cp2)
 324     if (toupper (*cp) != *cp2)
 325       FAIL ("toupper ('%c') != '%c'", *cp, *cp2);
 326   for (cp = upper; *cp != '\0'; ++cp)
 327     if (toupper (*cp) != *cp)
 328       FAIL ("toupper ('%c') != '%c'", *cp, *cp);
 329   for (cp = digits; *cp != '\0'; ++cp)
 330     if (toupper (*cp) != *cp)
 331       FAIL ("toupper ('%c') != '%c'", *cp, *cp);
 332   for (cp = cntrl; *cp != '\0'; ++cp)
 333     if (toupper (*cp) != *cp)
 334       FAIL ("toupper ('\\x%02x') != '\\x%02x'", *cp, *cp);
 335
 336
 337   /* Now some locale specific tests.  */
 338   while (! feof (stdin))
 339     {
 340       unsigned char *inp;
 341       unsigned char *resp;
 342
 343       if (getline (&inpline, &inplinelen, stdin) <= 0
 344           || getline (&resline, &reslinelen, stdin) <= 0)
 345         break;
 346
 347       inp = (unsigned char *) strchr (inpline, '\n');
 348       if (inp != NULL)
 349         *inp = '\0';
 350       resp = (unsigned char *) strchr (resline, '\n');
 351       if (resp != NULL)
 352         *resp = '\0';
 353
 354       inp = (unsigned char *) inpline;
 355       while (*inp != ' ' && *inp != '\t' && *inp && *inp != '\n'
 356              && *inp != '\0')
 357         ++inp;
 358
 359       if (*inp == '\0')
 360         {
 361           printf ("line \"%s\" is without content\n", inpline);
 362           continue;
 363         }
 364       *inp++ = '\0';
 365       while (*inp == ' ' || *inp == '\t')
 366         ++inp;
 367
 368       /* Try all classes.  */
 369       for (n = 0; n < nclasses; ++n)
 370         if (strcmp (inpline, classes[n].name) == 0)
 371           break;
 372
 373       resp = (unsigned char *) resline;
 374       while (*resp == ' ' || *resp == '\t')
 375         ++resp;
 376
 377       if (strlen ((char *) inp) != strlen ((char *) resp))
 378         {
 379           printf ("lines \"%.20s\"... and \"%.20s\" have not the same length\n",
 380                   inp, resp);
 381           continue;
 382         }
 383
 384       if (n < nclasses)
 385         {
 386           if (strspn ((char *) resp, "01") != strlen ((char *) resp))
 387             {
 388               printf ("result string \"%s\" malformed\n", resp);
 389               continue;
 390             }
 391
 392           printf ("  Locale-specific tests for `%s'\n", inpline);
 393
 394           while (*inp != '\0' && *inp != '\n')
 395             {
 396               if (((__ctype_b[(unsigned int) *inp] & classes[n].mask) != 0)
 397                   != (*resp != '0'))
 398                 {
 399                   printf ("    is%s('%c' = '\\x%02x') %s true\n", inpline,
 400                           *inp, *inp, *resp == '1' ? "not" : "is");
 401                   ++errors;
 402                 }
 403               ++inp;
 404               ++resp;
 405             }
 406         }
 407       else if (strcmp (inpline, "tolower") == 0)
 408         {
 409           while (*inp != '\0')
 410             {
 411               if (tolower (*inp) != *resp)
 412                 {
 413                   printf ("    tolower('%c' = '\\x%02x') != '%c'\n",
 414                           *inp, *inp, *resp);
 415                   ++errors;
 416                 }
 417               ++inp;
 418               ++resp;
 419             }
 420         }
 421       else if (strcmp (inpline, "toupper") == 0)
 422         {
 423           while (*inp != '\0')
 424             {
 425               if (toupper (*inp) != *resp)
 426                 {
 427                   printf ("    toupper('%c' = '\\x%02x') != '%c'\n",
 428                           *inp, *inp, *resp);
 429                   ++errors;
 430                 }
 431               ++inp;
 432               ++resp;
 433             }
 434         }
 435       else
 436         printf ("\"%s\": unknown class or map\n", inpline);
 437     }
 438
 439
 440   if (errors != 0)
 441     {
 442       printf ("  %d error%s for `%s' locale\n\n\n", errors,
 443               errors == 1 ? "" : "s", setlocale (LC_ALL, NULL));
 444       return 1;
 445     }
 446
 447   printf ("  No errors for `%s' locale\n\n\n", setlocale (LC_ALL, NULL));
 448   return 0;
 449 }