1 /*
2 *
3 * Copyright (c) 2004
4 * John Maddock
5 *
6 * Use, modification and distribution are subject to the
7 * Boost Software License, Version 1.0. (See accompanying file
8 * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
9 *
10 */
11
12 /*
13 * LOCATION: see http://www.boost.org for most recent version.
14 * FILE captures_test.cpp
15 * VERSION see <boost/version.hpp>
16 * DESCRIPTION: Basic tests for additional captures information.
17 */
18
19 #include <boost/regex.hpp>
20 #include <boost/detail/lightweight_main.hpp>
21 #include "../test_macros.hpp"
22 #include <boost/array.hpp>
23 #include <cstring>
24
25 #ifdef BOOST_HAS_ICU
26 #include <boost/regex/icu.hpp>
27 #endif
28
29 #define ARRAY_SIZE(x) (sizeof(x) / sizeof(x[0]))
30
31 template <int N>
array_size(const char * (& p)[N])32 size_t array_size(const char* (&p)[N])
33 {
34 for(size_t i = 0; i < N; ++i)
35 if(p[i] == 0)
36 return i;
37 return N;
38 }
39
make_wstring(const char * p)40 std::wstring make_wstring(const char* p)
41 {
42 return std::wstring(p, p + std::strlen(p));
43 }
44
45 #ifdef __sgi
46 template <class T>
test_captures(const std::string & regx,const std::string & text,const T & expected)47 void test_captures(const std::string& regx, const std::string& text, const T& expected)
48 #else
49 template <class T>
50 void test_captures(const std::string& regx, const std::string& text, T& expected)
51 #endif
52 {
53 boost::regex e(regx);
54 boost::smatch what;
55 if(boost::regex_match(text, what, e, boost::match_extra))
56 {
57 unsigned i, j;
58 #ifndef __sgi
59 // strange type deduction causes this test to fail on SGI:
60 BOOST_CHECK(what.size() == ARRAY_SIZE(expected));
61 #endif
62 for(i = 0; i < what.size(); ++i)
63 {
64 BOOST_CHECK(what.captures(i).size() == array_size(expected[i]));
65 for(j = 0; j < what.captures(i).size(); ++j)
66 {
67 BOOST_CHECK(what.captures(i)[j] == expected[i][j]);
68 }
69 }
70 }
71
72 #if !defined(BOOST_NO_WREGEX)
73
74 std::wstring wre(regx.begin(), regx.end());
75 std::wstring wtext(text.begin(), text.end());
76 boost::wregex we(wre);
77 boost::wsmatch wwhat;
78 if(boost::regex_match(wtext, wwhat, we, boost::match_extra))
79 {
80 unsigned i, j;
81 #ifndef __sgi
82 // strange type deduction causes this test to fail on SGI:
83 BOOST_CHECK(wwhat.size() == ARRAY_SIZE(expected));
84 #endif
85 for(i = 0; i < wwhat.size(); ++i)
86 {
87 BOOST_CHECK(wwhat.captures(i).size() == array_size(expected[i]));
88 for(j = 0; j < wwhat.captures(i).size(); ++j)
89 {
90 BOOST_CHECK(wwhat.captures(i)[j] == make_wstring(expected[i][j]));
91 }
92 }
93 }
94
95 #endif
96
97 #ifdef BOOST_HAS_ICU
98 boost::u32regex ure = boost::make_u32regex(regx);
99 what = boost::smatch();
100 if(boost::u32regex_match(text, what, ure, boost::match_extra))
101 {
102 unsigned i, j;
103 #ifndef __sgi
104 // strange type deduction causes this test to fail on SGI:
105 BOOST_CHECK(what.size() == ARRAY_SIZE(expected));
106 #endif
107 for(i = 0; i < what.size(); ++i)
108 {
109 BOOST_CHECK(what.captures(i).size() == array_size(expected[i]));
110 for(j = 0; j < what.captures(i).size(); ++j)
111 {
112 BOOST_CHECK(what.captures(i)[j] == expected[i][j]);
113 }
114 }
115 }
116 #endif
117 }
118
cpp_main(int,char * [])119 int cpp_main(int , char* [])
120 {
121 typedef const char* pchar;
122 pchar e1[4][5] =
123 {
124 { "aBBcccDDDDDeeeeeeee", },
125 { "a", "BB", "ccc", "DDDDD", "eeeeeeee", },
126 { "a", "ccc", "eeeeeeee", },
127 { "BB", "DDDDD", },
128 };
129 test_captures("(([[:lower:]]+)|([[:upper:]]+))+", "aBBcccDDDDDeeeeeeee", e1);
130 pchar e2[4][2] =
131 {
132 { "abd" },
133 { "b", "" },
134 { "" },
135 };
136 test_captures("a(b+|((c)*))+d", "abd", e2);
137 pchar e3[3][1] =
138 {
139 { "abcbar" },
140 { "abc" },
141 };
142 test_captures("(.*)bar|(.*)bah", "abcbar", e3);
143 pchar e4[3][1] =
144 {
145 { "abcbah" },
146 { 0, },
147 { "abc" },
148 };
149 test_captures("(.*)bar|(.*)bah", "abcbah", e4);
150 pchar e5[2][16] =
151 {
152 { "now is the time for all good men to come to the aid of the party" },
153 { "now", "is", "the", "time", "for", "all", "good", "men", "to", "come", "to", "the", "aid", "of", "the", "party" },
154 };
155 test_captures("^(?:(\\w+)|(?>\\W+))*$", "now is the time for all good men to come to the aid of the party", e5);
156 pchar e6[2][16] =
157 {
158 { "now is the time for all good men to come to the aid of the party" },
159 { "now", "is", "the", "time", "for", "all", "good", "men", "to", "come", "to", "the", "aid", "of", "the", "party" },
160 };
161 test_captures("^(?>(\\w+)\\W*)*$", "now is the time for all good men to come to the aid of the party", e6);
162 pchar e7[4][14] =
163 {
164 { "now is the time for all good men to come to the aid of the party" },
165 { "now" },
166 { "is", "the", "time", "for", "all", "good", "men", "to", "come", "to", "the", "aid", "of", "the" },
167 { "party" },
168 };
169 test_captures("^(\\w+)\\W+(?>(\\w+)\\W+)*(\\w+)$", "now is the time for all good men to come to the aid of the party", e7);
170 pchar e8[5][9] =
171 {
172 { "now is the time for all good men to come to the aid of the party" } ,
173 { "now" },
174 { "is", "for", "men", "to", "of" },
175 { "the", "time", "all", "good", "to", "come", "the", "aid", "the" },
176 { "party" },
177 };
178 test_captures("^(\\w+)\\W+(?>(\\w+)\\W+(?:(\\w+)\\W+){0,2})*(\\w+)$", "now is the time for all good men to come to the aid of the party", e8);
179 return 0;
180 }
181
182