tests/tst_stringprep.c

   1 /* tst_stringprep.c --- Self tests for stringprep().
   2  * Copyright (C) 2002, 2003, 2004  Simon Josefsson
   3  *
   4  * This file is part of GNU Libidn.
   5  *
   6  * GNU Libidn is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * GNU Libidn 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
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with GNU Libidn; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
  19  *
  20  */
  21
  22 #ifdef HAVE_CONFIG_H
  23 # include "config.h"
  24 #endif
  25
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <stdarg.h>
  29 #include <string.h>
  30
  31 #include <stringprep.h>
  32
  33 #include "utils.h"
  34
  35 struct stringprep
  36 {
  37   const char *comment;
  38   const char *in;
  39   const char *out;
  40   const char *profile;
  41   int flags;
  42   int rc;
  43 };
  44
  45 const struct stringprep strprep[] = {
  46   {"Map to nothing",
  47    "foo\xC2\xAD\xCD\x8F\xE1\xA0\x86\xE1\xA0\x8B"
  48    "bar" "\xE2\x80\x8B\xE2\x81\xA0" "baz\xEF\xB8\x80\xEF\xB8\x88"
  49    "\xEF\xB8\x8F\xEF\xBB\xBF", "foobarbaz"},
  50   {"Case folding ASCII U+0043 U+0041 U+0046 U+0045", "CAFE", "cafe"},
  51   {"Case folding 8bit U+00DF (german sharp s)", "\xC3\x9F", "ss"},
  52   {"Case folding U+0130 (turkish capital I with dot)",
  53    "\xC4\xB0", "i\xcc\x87"},
  54   {"Case folding multibyte U+0143 U+037A",
  55    "\xC5\x83\xCD\xBA", "\xC5\x84 \xCE\xB9"},
  56   {"Case folding U+2121 U+33C6 U+1D7BB",
  57    "\xE2\x84\xA1\xE3\x8F\x86\xF0\x9D\x9E\xBB",
  58    "telc\xE2\x88\x95" "kg\xCF\x83"},
  59   {"Normalization of U+006a U+030c U+00A0 U+00AA",
  60    "\x6A\xCC\x8C\xC2\xA0\xC2\xAA", "\xC7\xB0 a"},
  61   {"Case folding U+1FB7 and normalization",
  62    "\xE1\xBE\xB7", "\xE1\xBE\xB6\xCE\xB9"},
  63   {"Self-reverting case folding U+01F0 and normalization",
  64    "\xC7\xB0", "\xC7\xB0"},
  65   {"Self-reverting case folding U+0390 and normalization",
  66    "\xCE\x90", "\xCE\x90"},
  67   {"Self-reverting case folding U+03B0 and normalization",
  68    "\xCE\xB0", "\xCE\xB0"},
  69   {"Self-reverting case folding U+1E96 and normalization",
  70    "\xE1\xBA\x96", "\xE1\xBA\x96"},
  71   {"Self-reverting case folding U+1F56 and normalization",
  72    "\xE1\xBD\x96", "\xE1\xBD\x96"},
  73   {"ASCII space character U+0020", "\x20", "\x20"},
  74   {"Non-ASCII 8bit space character U+00A0", "\xC2\xA0", "\x20"},
  75   {"Non-ASCII multibyte space character U+1680",
  76    "\xE1\x9A\x80", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
  77   {"Non-ASCII multibyte space character U+2000", "\xE2\x80\x80", "\x20"},
  78   {"Zero Width Space U+200b", "\xE2\x80\x8b", ""},
  79   {"Non-ASCII multibyte space character U+3000", "\xE3\x80\x80", "\x20"},
  80   {"ASCII control characters U+0010 U+007F", "\x10\x7F", "\x10\x7F"},
  81   {"Non-ASCII 8bit control character U+0085",
  82    "\xC2\x85", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
  83   {"Non-ASCII multibyte control character U+180E",
  84    "\xE1\xA0\x8E", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
  85   {"Zero Width No-Break Space U+FEFF", "\xEF\xBB\xBF", ""},
  86   {"Non-ASCII control character U+1D175",
  87    "\xF0\x9D\x85\xB5", NULL, "Nameprep", 0,
  88    STRINGPREP_CONTAINS_PROHIBITED},
  89   {"Plane 0 private use character U+F123",
  90    "\xEF\x84\xA3", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
  91   {"Plane 15 private use character U+F1234",
  92    "\xF3\xB1\x88\xB4", NULL, "Nameprep", 0,
  93    STRINGPREP_CONTAINS_PROHIBITED},
  94   {"Plane 16 private use character U+10F234",
  95    "\xF4\x8F\x88\xB4", NULL, "Nameprep", 0,
  96    STRINGPREP_CONTAINS_PROHIBITED},
  97   {"Non-character code point U+8FFFE",
  98    "\xF2\x8F\xBF\xBE", NULL, "Nameprep", 0,
  99    STRINGPREP_CONTAINS_PROHIBITED},
 100   {"Non-character code point U+10FFFF",
 101    "\xF4\x8F\xBF\xBF", NULL, "Nameprep", 0,
 102    STRINGPREP_CONTAINS_PROHIBITED},
 103   {"Surrogate code U+DF42",
 104    "\xED\xBD\x82", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 105   {"Non-plain text character U+FFFD",
 106    "\xEF\xBF\xBD", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 107   {"Ideographic description character U+2FF5",
 108    "\xE2\xBF\xB5", NULL, "Nameprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 109   {"Display property character U+0341", "\xCD\x81", "\xCC\x81"},
 110   {"Left-to-right mark U+200E",
 111    "\xE2\x80\x8E", "\xCC\x81", "Nameprep", 0,
 112    STRINGPREP_CONTAINS_PROHIBITED},
 113   {"Deprecated U+202A", "\xE2\x80\xAA", "\xCC\x81", "Nameprep", 0,
 114    STRINGPREP_CONTAINS_PROHIBITED},
 115   {"Language tagging character U+E0001",
 116    "\xF3\xA0\x80\x81", "\xCC\x81", "Nameprep", 0,
 117    STRINGPREP_CONTAINS_PROHIBITED},
 118   {"Language tagging character U+E0042",
 119    "\xF3\xA0\x81\x82", NULL, "Nameprep", 0,
 120    STRINGPREP_CONTAINS_PROHIBITED},
 121   {"Bidi: RandALCat character U+05BE and LCat characters",
 122    "foo\xD6\xBE" "bar", NULL, "Nameprep", 0,
 123    STRINGPREP_BIDI_BOTH_L_AND_RAL},
 124   {"Bidi: RandALCat character U+FD50 and LCat characters",
 125    "foo\xEF\xB5\x90" "bar", NULL, "Nameprep", 0,
 126    STRINGPREP_BIDI_BOTH_L_AND_RAL},
 127   {"Bidi: RandALCat character U+FB38 and LCat characters",
 128    "foo\xEF\xB9\xB6" "bar", "foo \xd9\x8e" "bar"},
 129   {"Bidi: RandALCat without trailing RandALCat U+0627 U+0031",
 130    "\xD8\xA7\x31", NULL, "Nameprep", 0, STRINGPREP_BIDI_LEADTRAIL_NOT_RAL},
 131   {"Bidi: RandALCat character U+0627 U+0031 U+0628",
 132    "\xD8\xA7\x31\xD8\xA8", "\xD8\xA7\x31\xD8\xA8"},
 133   {"Unassigned code point U+E0002",
 134    "\xF3\xA0\x80\x82", NULL, "Nameprep", STRINGPREP_NO_UNASSIGNED,
 135    STRINGPREP_CONTAINS_UNASSIGNED},
 136   {"Larger test (shrinking)",
 137    "X\xC2\xAD\xC3\x9F\xC4\xB0\xE2\x84\xA1\x6a\xcc\x8c\xc2\xa0\xc2"
 138    "\xaa\xce\xb0\xe2\x80\x80", "xssi\xcc\x87" "tel\xc7\xb0 a\xce\xb0 ",
 139    "Nameprep"},
 140   {"Larger test (expanding)",
 141    "X\xC3\x9F\xe3\x8c\x96\xC4\xB0\xE2\x84\xA1\xE2\x92\x9F\xE3\x8c\x80",
 142    "xss\xe3\x82\xad\xe3\x83\xad\xe3\x83\xa1\xe3\x83\xbc\xe3\x83\x88"
 143    "\xe3\x83\xab" "i\xcc\x87" "tel\x28" "d\x29\xe3\x82\xa2\xe3\x83\x91"
 144    "\xe3\x83\xbc\xe3\x83\x88"},
 145   {"Test of prohibited ASCII character U+0020",
 146    "\x20", NULL, "Nodeprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 147   {"Test of NFKC U+00A0 and prohibited character U+0020",
 148    "\xC2\xA0", NULL, "Nodeprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 149   {"Case map + normalization", "\xC2\xB5", "\xCE\xBC", "Nameprep"},
 150   /* The rest are rather non-interesting, but no point in removing
 151      working test cases... */
 152   {"case_nonfkc", "\xC2\xB5", "\xCE\xBC", "Nameprep", STRINGPREP_NO_NFKC,
 153    STRINGPREP_FLAG_ERROR},
 154   {"NFKC test", "\xC2\xAA", "\x61", "Nameprep"},
 155   {"nameprep, exposed a bug in libstringprep 0.0.5",
 156    "\xC2\xAA\x0A", "\x61\x0A"},
 157   {"unassigned code point U+0221", "\xC8\xA1", "\xC8\xA1", "Nameprep"},
 158   {"Unassigned code point U+0221",
 159    "\xC8\xA1", NULL, "Nameprep", STRINGPREP_NO_UNASSIGNED,
 160    STRINGPREP_CONTAINS_UNASSIGNED},
 161   {"Unassigned code point U+0236", "\xC8\xB6", "\xC8\xB6", "Nameprep"},
 162   {"unassigned code point U+0236",
 163    "\xC8\xB6", NULL, "Nameprep", STRINGPREP_NO_UNASSIGNED,
 164    STRINGPREP_CONTAINS_UNASSIGNED},
 165   {"bidi both RandALCat and LCat  U+0627 U+00AA U+0628",
 166    "\xD8\xA7\xC2\xAA\xD8\xA8", NULL, "Nameprep", 0,
 167    STRINGPREP_BIDI_BOTH_L_AND_RAL},
 168   /* XMPP */
 169   {"XMPP node profile prohibited output",
 170    "foo@bar", NULL, "Nodeprep", 0, STRINGPREP_CONTAINS_PROHIBITED},
 171   {"XMPP resource profile on same string should work though",
 172    "foo@bar", "foo@bar", "Resourceprep"},
 173   /* iSCSI */
 174   {"iSCSI 1", "Example-Name", "example-name", "iSCSI"},
 175   {"iSCSI 2", "O+o", NULL, "iSCSI", 0, STRINGPREP_CONTAINS_PROHIBITED},
 176   {"iSCSI 3", "\x01", NULL, "iSCSI", 0, STRINGPREP_CONTAINS_PROHIBITED},
 177   {"iSCSI 4", "\xE3\x80\x82", NULL, "iSCSI", 0,
 178    STRINGPREP_CONTAINS_PROHIBITED},
 179   {"iSCSI 5", "\xE2\xBF\xB5", NULL, "iSCSI", 0,
 180    STRINGPREP_CONTAINS_PROHIBITED},
 181   {"SASL profile", "Example\xC2\xA0" "Name", "Example Name", "SASLprep"},
 182   /* SASL trace */
 183   {"SASL ANONYMOUS plain mechanism", "simon@josefsson.org",
 184    "simon@josefsson.org", "plain"},
 185   {"SASLprep 1 old", "x\xC2\xADy", "xy", "SASLprep"},
 186   {"SASLprep 4 old", "\xE2\x85\xA3", "IV", "SASLprep"},
 187   /* SASLprep test vectors. */
 188   {"SASLprep 1 SOFT HYPHEN mapped to nothing", "I\xC2\xADX", "IX",
 189    "SASLprep"},
 190   {"SASLprep 2 no transformation", "user", "user", "SASLprep"},
 191   {"SASLprep 3 case preserved, will not match #2", "USER", "USER",
 192    "SASLprep"},
 193   {"SASLprep 4 output is NFKC, input in ISO 8859-1", "\xC2\xAA", "a",
 194    "SASLprep"},
 195   {"SASLprep 5 output is NFKC, will match #1", "\xE2\x85\xA8", "IX",
 196    "SASLprep"},
 197   {"SASLprep 6 Error - prohibited character", "\x07", NULL, "SASLprep",
 198    0, STRINGPREP_CONTAINS_PROHIBITED},
 199   {"SASLprep 7 Error - bidirectional check", "\xD8\xA7" "1", NULL, "SASLprep",
 200    0, STRINGPREP_BIDI_LEADTRAIL_NOT_RAL}
 201 };
 202
 203 void
 204 doit (void)
 205 {
 206   char *p;
 207   int rc;
 208   size_t i;
 209
 210   if (!stringprep_check_version (STRINGPREP_VERSION))
 211     fail ("stringprep_check_version() failed\n");
 212
 213   for (i = 0; i < sizeof (strprep) / sizeof (strprep[0]); i++)
 214     {
 215       if (debug)
 216         printf ("STRINGPREP entry %d\n", i);
 217
 218       if (debug)
 219         {
 220           printf ("flags: %d\n", strprep[i].flags);
 221
 222           printf ("in: ");
 223           escapeprint (strprep[i].in, strlen (strprep[i].in));
 224           hexprint (strprep[i].in, strlen (strprep[i].in));
 225           binprint (strprep[i].in, strlen (strprep[i].in));
 226         }
 227
 228       {
 229         uint32_t *l;
 230         char *x;
 231         l = stringprep_utf8_to_ucs4 (strprep[i].in, -1, NULL);
 232         x = stringprep_ucs4_to_utf8 (l, -1, NULL, NULL);
 233         free (l);
 234
 235         if (strcmp (strprep[i].in, x) != 0)
 236           {
 237             fail ("bad UTF-8 in entry %d\n", i);
 238             if (debug)
 239               {
 240                 puts ("expected:");
 241                 escapeprint (strprep[i].in, strlen (strprep[i].in));
 242                 hexprint (strprep[i].in, strlen (strprep[i].in));
 243                 puts ("computed:");
 244                 escapeprint (x, strlen (x));
 245                 hexprint (x, strlen (x));
 246               }
 247           }
 248
 249         free (x);
 250       }
 251       rc = stringprep_profile (strprep[i].in, &p,
 252                                strprep[i].profile ?
 253                                strprep[i].profile :
 254                                "Nameprep", strprep[i].flags);
 255       if (rc != strprep[i].rc)
 256         {
 257           fail ("stringprep() entry %d failed: %d\n", i, rc);
 258           if (debug)
 259             printf ("FATAL\n");
 260           if (rc == STRINGPREP_OK)
 261             free (p);
 262           continue;
 263         }
 264
 265       if (debug && rc == STRINGPREP_OK)
 266         {
 267           printf ("out: ");
 268           escapeprint (p, strlen (p));
 269           hexprint (p, strlen (p));
 270           binprint (p, strlen (p));
 271
 272           printf ("expected out: ");
 273           escapeprint (strprep[i].out, strlen (strprep[i].out));
 274           hexprint (strprep[i].out, strlen (strprep[i].out));
 275           binprint (strprep[i].out, strlen (strprep[i].out));
 276         }
 277       else if (debug)
 278         printf ("returned %d expected %d\n", rc, strprep[i].rc);
 279
 280       if (rc == STRINGPREP_OK)
 281         {
 282           if (strlen (strprep[i].out) != strlen (p) ||
 283               memcmp (strprep[i].out, p, strlen (p)) != 0)
 284             {
 285               fail ("stringprep() entry %d failed\n", i);
 286               if (debug)
 287                 printf ("ERROR\n");
 288             }
 289           else if (debug)
 290             printf ("OK\n\n");
 291
 292           free (p);
 293         }
 294       else if (debug)
 295         printf ("OK\n\n");
 296     }
 297
 298 #if 0
 299   {
 300     char p[20];
 301     memset (p, 0, 10);
 302     stringprep_unichar_to_utf8 (0x00DF, p);
 303     hexprint (p, strlen (p));
 304     puts ("");
 305   }
 306 #endif
 307 }