2 This file is part of GNUnet
3 (C) 2012 Christian Grothoff (and other contributing authors)
5 GNUnet is free software; you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published
7 by the Free Software Foundation; either version 3, or (at your
8 option) any later version.
10 GNUnet is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with GNUnet; see the file COPYING. If not, write to the
17 Free Software Foundation, Inc., 59 Temple Place - Suite 330,
18 Boston, MA 02111-1307, USA.
21 * @file regex/test_regex_eval_api.c
22 * @brief test for regex.c
23 * @author Maximilian Szengel
28 #include "gnunet_regex_lib.h"
29 #include "regex_internal.h"
37 struct Regex_String_Pair
42 enum Match_Result expected_results[20];
47 * Random regex test. Generate a random regex as well as 'str_count' strings to
48 * match it against. Will match using GNUNET_REGEX implementation and compare
49 * the result to glibc regex result. 'rx_length' has to be smaller then 'max_str_len'.
51 * @param rx_length length of the regular expression.
52 * @param max_str_len maximum length of the random strings.
53 * @param str_count number of generated random strings.
55 * @return 0 on success, non 0 otherwise.
58 test_random (unsigned int rx_length, unsigned int max_str_len,
59 unsigned int str_count)
67 struct GNUNET_REGEX_Automaton *dfa;
69 regmatch_t matchptr[1];
73 char *canonical_regex;
75 // At least one string is needed for matching
76 GNUNET_assert (str_count > 0);
77 // The string should be at least as long as the regex itself
78 GNUNET_assert (max_str_len >= rx_length);
80 // Generate random regex and a string that matches the regex
81 matching_str = GNUNET_malloc (rx_length + 1);
82 rand_rx = GNUNET_REGEX_generate_random_regex (rx_length, matching_str);
86 for (i = 0; i < str_count; i++)
90 matching_str = GNUNET_REGEX_generate_random_string (max_str_len);
91 str_len = strlen (matching_str);
94 // Match string using DFA
95 dfa = GNUNET_REGEX_construct_dfa (rand_rx, strlen (rand_rx));
98 GNUNET_log (GNUNET_ERROR_TYPE_ERROR, "Constructing DFA failed\n");
102 eval = GNUNET_REGEX_eval (dfa, matching_str);
103 canonical_regex = GNUNET_strdup (GNUNET_REGEX_get_canonical_regex (dfa));
104 GNUNET_REGEX_automaton_destroy (dfa);
106 // Match string using glibc regex
107 if (0 != regcomp (&rx, rand_rx, REG_EXTENDED))
109 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
110 "Could not compile regex using regcomp\n");
114 eval_check = regexec (&rx, matching_str, 1, matchptr, 0);
117 // We only want to match the whole string, because that's what our DFA does, too.
118 if (eval_check == 0 &&
119 (matchptr[0].rm_so != 0 || matchptr[0].rm_eo != strlen (matching_str)))
122 // Match canonical regex
123 if (0 != regcomp (&rx, canonical_regex, REG_EXTENDED))
125 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
126 "Could not compile regex using regcomp: %s\n",
131 eval_canonical = regexec (&rx, matching_str, 1, matchptr, 0);
133 GNUNET_free (canonical_regex);
136 if (eval_check != eval)
138 regerror (eval_check, &rx, error, sizeof error);
139 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
140 "Unexpected result:\nregex: %s\nstring: %s\ngnunet regex: %i\nglibc regex: %i\nglibc error: %s\n\n",
141 rand_rx, matching_str, eval, eval_check, error);
145 GNUNET_free (matching_str);
148 GNUNET_free (rand_rx);
154 * Automaton test that compares the result of matching regular expression 'rx'
155 * with the strings and expected results in 'rxstr' with the result of matching
156 * the same strings with glibc regex.
158 * @param a automaton.
159 * @param rx compiled glibc regex.
160 * @param rxstr regular expression and strings with expected results to match against.
162 * @return 0 on successfull, non 0 otherwise
165 test_automaton (struct GNUNET_REGEX_Automaton *a, regex_t * rx,
166 struct Regex_String_Pair *rxstr)
172 regmatch_t matchptr[1];
177 GNUNET_log (GNUNET_ERROR_TYPE_ERROR, "Automaton was NULL\n");
183 for (i = 0; i < rxstr->string_count; i++)
185 eval = GNUNET_REGEX_eval (a, rxstr->strings[i]);
186 eval_check = regexec (rx, rxstr->strings[i], 1, matchptr, 0);
188 // We only want to match the whole string, because that's what our DFA does, too.
189 if (eval_check == 0 &&
190 (matchptr[0].rm_so != 0 ||
191 matchptr[0].rm_eo != strlen (rxstr->strings[i])))
194 if ((rxstr->expected_results[i] == match && (0 != eval || 0 != eval_check))
195 || (rxstr->expected_results[i] == nomatch &&
196 (0 == eval || 0 == eval_check)))
199 regerror (eval_check, rx, error, sizeof error);
200 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
201 "Unexpected result:\nregex: %s\ncanonical_regex: %s\nstring: %s\nexpected result: %i\n"
202 "gnunet regex: %i\nglibc regex: %i\nglibc error: %s\nrm_so: %i\nrm_eo: %i\n\n",
203 rxstr->regex, GNUNET_REGEX_get_canonical_regex (a),
204 rxstr->strings[i], rxstr->expected_results[i], eval,
205 eval_check, error, matchptr[0].rm_so, matchptr[0].rm_eo);
212 main (int argc, char *argv[])
214 GNUNET_log_setup ("test-regex",
222 struct GNUNET_REGEX_Automaton *a;
230 struct Regex_String_Pair rxstr[16] = {
232 {"ababcd", "abab", "aabcd", "a", "abb"},
233 {match, nomatch, match, match, nomatch}},
234 {"ab(c|d)+c*(a(b|c)d)+", 5,
235 {"abcdcdcdcdddddabd", "abcd", "abcddddddccccccccccccccccccccccccabdacdabd",
236 "abccccca", "abcdcdcdccdabdabd"},
237 {match, nomatch, match, nomatch, match}},
238 {"ab+c*(a(bx|c)d)+", 5,
239 {"abcdcdcdcdddddabd", "abcd", "abcddddddccccccccccccccccccccccccabdacdabd",
240 "abccccca", "abcdcdcdccdabdabd"},
241 {nomatch, nomatch, nomatch, nomatch, nomatch}},
242 {"a+X*y+c|p|R|Z*K*y*R+w|Y*6+n+h*k*w+V*F|W*B*e*", 1,
243 {"kaXycQepRZKyRwY6nhkwVFWBegNVtLPj39XhJJ6bEifRSZRYZg"},
245 {"k|a+X*y+c|Q*e|p|R|Z*K*y*R+w|Y*6+n+h*k*w+V*F|W*B*e*g|N+V|t+L|P*j*3*9+X*h*J|J*6|b|E*i*f*R+S|Z|R|Y*Z|g*", 1,
246 {"kaXycQepRZKyRwY6nhkwVFWBegNVtLPj39XhJJ6bEifRSZRYZg"},
248 {"F?W+m+2*6*c*s|P?U?a|B|y*i+t+A|V|6*C*7*e?Z*n*i|J?5+g?W*V?7*j?p?1|r?B?C+E+3+6*i+W*P?K?0|D+7?y*m+3?g?K?", 1,
249 {"osfjsodfonONONOnosndfsdnfsd"},
251 {"V|M*o?x*p*d+h+b|E*m?h?Y*E*O?W*W*P+o?Z+H*M|I*q+C*a+5?5*9|b?z|G*y*k?R|p+u|8*h?B+l*H|e|L*O|1|F?v*0?5|C+", 1,
252 {"VMoxpdhbEmhYEOWWPoZHMIqCa559bzGykRpu8hBlHeLO1Fv05C"},
255 {"", "bla", "blabla", "bl", "la", "b", "l", "a"},
256 {match, match, match, nomatch, nomatch, nomatch, nomatch, nomatch}},
257 {"ab(c|d)+c*(a(b|c)+d)+(bla)(bla)*", 8,
258 {"ab", "abcabdbla", "abdcccccccccccabcbccdblablabla", "bl", "la", "b", "l",
260 {nomatch, match, match, nomatch, nomatch, nomatch, nomatch, nomatch}},
262 {"", "a", "aa", "aaa", "aaaa", "aaaaa"},
263 {nomatch, match, match, match, match, match}},
264 {"ab(c|d)+c*(a(b|c)+d)+(bla)+", 1,
268 {"b", "bb", "ac", "", "acb", "bacbacac", "acacac", "abc"},
269 {match, match, match, nomatch, match, match, match, nomatch}},
271 {"", "ab", "c", "abc", "ababcc", "acc", "abac"},
272 {nomatch, match, match, match, match, nomatch, nomatch}},
273 {"((j|2j)K|(j|2j)AK|(j|2j)(D|e|(j|2j)A(D|e))D*K)", 1,
274 {"", "2j2jADK", "j2jADK"},
275 {nomatch, match, match}},
276 {"((j|2j)K|(j|2j)(D|e|((j|2j)j|(j|2j)2j)A(D|e))D*K|(j|2j)AK)", 2,
277 {"", "2j2jjADK", "j2jADK"},
278 {nomatch, match, match}},
279 {"ab(c|d)+c*(a(b|c)d)+", 1,
288 for (i = 0; i < 16; i++)
290 if (0 != regcomp (&rx, rxstr[i].regex, REG_EXTENDED))
292 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
293 "Could not compile regex using regcomp()\n");
298 a = GNUNET_REGEX_construct_nfa (rxstr[i].regex, strlen (rxstr[i].regex));
299 check_nfa += test_automaton (a, &rx, &rxstr[i]);
300 GNUNET_REGEX_automaton_destroy (a);
303 a = GNUNET_REGEX_construct_dfa (rxstr[i].regex, strlen (rxstr[i].regex));
304 check_dfa += test_automaton (a, &rx, &rxstr[i]);
305 check_proof = GNUNET_strdup (GNUNET_REGEX_get_canonical_regex (a));
306 GNUNET_REGEX_automaton_destroy (a);
308 a = GNUNET_REGEX_construct_dfa (check_proof, strlen (check_proof));
309 check_dfa += test_automaton (a, &rx, &rxstr[i]);
310 GNUNET_REGEX_automaton_destroy (a);
312 GNUNET_log (GNUNET_ERROR_TYPE_DEBUG, "check_proof: %s\n", check_proof);
313 GNUNET_free_non_null (check_proof);
319 for (i = 0; i < 50; i++)
320 check_rand += test_random (100, 120, 20);
322 return check_nfa + check_dfa + check_rand;