]> git.saurik.com Git - wxWidgets.git/blob - tests/regex/regextest.cpp
Fix install_name_tool calls in OS X "make install".
[wxWidgets.git] / tests / regex / regextest.cpp
1 ///////////////////////////////////////////////////////////////////////////////
2 // Name: tests/regex/regex.cpp
3 // Purpose: Test the built-in regex lib and wxRegEx
4 // Author: Mike Wetherell
5 // Copyright: (c) 2004 Mike Wetherell
6 // Licence: wxWindows licence
7 ///////////////////////////////////////////////////////////////////////////////
8
9 //
10 // Notes:
11 //
12 // To run just one section, say wx_1, do this:
13 // test regex.wx_1
14 //
15 // To run all the regex tests:
16 // test regex
17 //
18 // Some tests must be skipped since they use features which we do not make
19 // available through wxRegEx. To see the list of tests that have been skipped
20 // turn on verbose logging, e.g.:
21 // test --verbose regex
22 //
23 // The tests here are for the builtin library, tests for wxRegEx in general
24 // should go in wxregex.cpp
25 //
26 // The tests are generated from Henry Spencer's reg.test, additional test
27 // can be added in wxreg.test. These test files are then turned into a C++
28 // include file 'regex.inc' (included below) using a script 'regex.pl'.
29 //
30
31 // For compilers that support precompilation, includes "wx/wx.h".
32 #include "testprec.h"
33
34 #ifdef __BORLANDC__
35 #pragma hdrstop
36 #endif
37
38 #if wxUSE_REGEX
39
40 // for all others, include the necessary headers
41 #ifndef WX_PRECOMP
42 #include "wx/wx.h"
43 #endif
44
45
46 // many of the tests are specific to the builtin regex lib, so only attempts
47 // to do them when using the builtin regex lib.
48 //
49 #ifdef wxHAS_REGEX_ADVANCED
50
51 #include "wx/regex.h"
52 #include <string>
53 #include <vector>
54
55 using CppUnit::Test;
56 using CppUnit::TestCase;
57 using CppUnit::TestSuite;
58 using CppUnit::Exception;
59
60 using std::string;
61 using std::vector;
62
63 ///////////////////////////////////////////////////////////////////////////////
64 // The test case - an instance represents a single test
65
66 class RegExTestCase : public TestCase
67 {
68 public:
69 // constructor - create a single testcase
70 RegExTestCase(
71 const string& name,
72 const char *mode,
73 const char *id,
74 const char *flags,
75 const char *pattern,
76 const char *data,
77 const vector<const char *>& expected);
78
79 protected:
80 // run this testcase
81 void runTest();
82
83 private:
84 // workers
85 wxString Conv(const char *str);
86 void parseFlags(const wxString& flags);
87 void doTest(int flavor);
88 static wxString quote(const wxString& arg);
89 const wxChar *convError() const { return wxT("<cannot convert>"); }
90
91 // assertions - adds some information about the test that failed
92 void fail(const wxString& msg) const;
93 void failIf(bool condition, const wxString& msg) const
94 { if (condition) fail(msg); }
95
96 // mode, id, flags, pattern, test data, expected results...
97 int m_mode;
98 wxString m_id;
99 wxString m_flags;
100 wxString m_pattern;
101 wxString m_data;
102 wxArrayString m_expected;
103
104 // the flag decoded
105 int m_compileFlags;
106 int m_matchFlags;
107 bool m_basic;
108 bool m_extended;
109 bool m_advanced;
110 };
111
112 // constructor - throws Exception on failure
113 //
114 RegExTestCase::RegExTestCase(
115 const string& name,
116 const char *mode,
117 const char *id,
118 const char *flags,
119 const char *pattern,
120 const char *data,
121 const vector<const char *>& expected)
122 :
123 TestCase(name),
124 m_mode(mode[0]),
125 m_id(Conv(id)),
126 m_flags(Conv(flags)),
127 m_pattern(Conv(pattern)),
128 m_data(Conv(data)),
129 m_compileFlags(0),
130 m_matchFlags(0),
131 m_basic(false),
132 m_extended(false),
133 m_advanced(false)
134 {
135 bool badconv = m_pattern == convError() || m_data == convError();
136 //RN: Removing the std:: here will break MSVC6 compilation
137 std::vector<const char *>::const_iterator it;
138
139 for (it = expected.begin(); it != expected.end(); ++it) {
140 m_expected.push_back(Conv(*it));
141 badconv = badconv || *m_expected.rbegin() == convError();
142 }
143
144 failIf(badconv, wxT("cannot convert to default character encoding"));
145
146 // the flags need further parsing...
147 parseFlags(m_flags);
148
149 #ifndef wxHAS_REGEX_ADVANCED
150 failIf(!m_basic && !m_extended, wxT("advanced regexs not available"));
151 #endif
152 }
153
154 int wxWcscmp(const wchar_t* s1, const wchar_t* s2)
155 {
156 size_t nLen1 = wxWcslen(s1);
157 size_t nLen2 = wxWcslen(s2);
158
159 if (nLen1 != nLen2)
160 return nLen1 - nLen2;
161
162 return memcmp(s1, s2, nLen1*sizeof(wchar_t));
163 }
164
165 // convert a string from UTF8 to the internal encoding
166 //
167 wxString RegExTestCase::Conv(const char *str)
168 {
169 const wxWCharBuffer wstr = wxConvUTF8.cMB2WC(str);
170 const wxWC2WXbuf buf = wxConvCurrent->cWC2WX(wstr);
171
172 if (!buf || wxWcscmp(wxConvCurrent->cWX2WC(buf), wstr) != 0)
173 return convError();
174
175 return buf;
176 }
177
178 // Parse flags
179 //
180 void RegExTestCase::parseFlags(const wxString& flags)
181 {
182 for ( wxString::const_iterator p = flags.begin(); p != flags.end(); ++p )
183 {
184 switch ( (*p).GetValue() ) {
185 // noop
186 case '-': break;
187
188 // we don't fully support these flags, but they don't stop us
189 // checking for success of failure of the match, so treat as noop
190 case 'A': case 'B': case 'E': case 'H':
191 case 'I': case 'L': case 'M': case 'N':
192 case 'P': case 'Q': case 'R': case 'S':
193 case 'T': case 'U': case '%':
194 break;
195
196 // match options
197 case '^': m_matchFlags |= wxRE_NOTBOL; break;
198 case '$': m_matchFlags |= wxRE_NOTEOL; break;
199 #if wxUSE_UNICODE
200 case '*': break;
201 #endif
202 // compile options
203 case '&': m_advanced = m_basic = true; break;
204 case 'b': m_basic = true; break;
205 case 'e': m_extended = true; break;
206 case 'i': m_compileFlags |= wxRE_ICASE; break;
207 case 'o': m_compileFlags |= wxRE_NOSUB; break;
208 case 'n': m_compileFlags |= wxRE_NEWLINE; break;
209 case 't': if (strchr("ep", m_mode)) break; // else fall through...
210
211 // anything else we must skip the test
212 default:
213 fail(wxString::Format(
214 wxT("requires unsupported flag '%c'"), *p));
215 }
216 }
217 }
218
219 // Try test for all flavours of expression specified
220 //
221 void RegExTestCase::runTest()
222 {
223 if (m_basic)
224 doTest(wxRE_BASIC);
225 if (m_extended)
226 doTest(wxRE_EXTENDED);
227 #ifdef wxHAS_REGEX_ADVANCED
228 if (m_advanced || (!m_basic && !m_extended))
229 doTest(wxRE_ADVANCED);
230 #endif
231 }
232
233 // Try the test for a single flavour of expression
234 //
235 void RegExTestCase::doTest(int flavor)
236 {
237 wxRegEx re(m_pattern, m_compileFlags | flavor);
238
239 // 'e' - test that the pattern fails to compile
240 if (m_mode == 'e') {
241 failIf(re.IsValid(), wxT("compile succeeded (should fail)"));
242 return;
243 }
244 failIf(!re.IsValid(), wxT("compile failed"));
245
246 bool matches = re.Matches(m_data, m_matchFlags);
247
248 // 'f' or 'p' - test that the pattern does not match
249 if (m_mode == 'f' || m_mode == 'p') {
250 failIf(matches, wxT("match succeeded (should fail)"));
251 return;
252 }
253
254 // otherwise 'm' or 'i' - test the pattern does match
255 failIf(!matches, wxT("match failed"));
256
257 if (m_compileFlags & wxRE_NOSUB)
258 return;
259
260 // check wxRegEx has correctly counted the number of subexpressions
261 wxString msg;
262 msg << wxT("GetMatchCount() == ") << re.GetMatchCount()
263 << wxT(", expected ") << m_expected.size();
264 failIf(m_expected.size() != re.GetMatchCount(), msg);
265
266 for (size_t i = 0; i < m_expected.size(); i++) {
267 wxString result;
268 size_t start, len;
269
270 msg.clear();
271 msg << wxT("wxRegEx::GetMatch failed for match ") << i;
272 failIf(!re.GetMatch(&start, &len, i), msg);
273
274 // m - check the match returns the strings given
275 if (m_mode == 'm')
276 {
277 if (start < INT_MAX)
278 result = m_data.substr(start, len);
279 else
280 result = wxT("");
281 }
282
283 // i - check the match returns the offsets given
284 else if (m_mode == 'i')
285 {
286 if (start > INT_MAX)
287 result = wxT("-1 -1");
288 else if (start + len > 0)
289 result << start << wxT(" ") << start + len - 1;
290 else
291 result << start << wxT(" -1");
292 }
293
294 msg.clear();
295 msg << wxT("match(") << i << wxT(") == ") << quote(result)
296 << wxT(", expected == ") << quote(m_expected[i]);
297 failIf(result != m_expected[i], msg);
298 }
299 }
300
301 // assertion - adds some information about the test that failed
302 //
303 void RegExTestCase::fail(const wxString& msg) const
304 {
305 wxString str;
306 wxArrayString::const_iterator it;
307
308 str << (wxChar)m_mode << wxT(" ") << m_id << wxT(" ") << m_flags << wxT(" ")
309 << quote(m_pattern) << wxT(" ") << quote(m_data);
310
311 for (it = m_expected.begin(); it != m_expected.end(); ++it)
312 str << wxT(" ") << quote(*it);
313
314 if (str.length() > 77)
315 str = str.substr(0, 74) + wxT("...");
316
317 str << wxT("\n ") << msg;
318
319 // no lossy convs so using utf8
320 CPPUNIT_FAIL(string(str.mb_str(wxConvUTF8)));
321 }
322
323 // quote a string so that it can be displayed (static)
324 //
325 wxString RegExTestCase::quote(const wxString& arg)
326 {
327 const wxChar *needEscape = wxT("\a\b\t\n\v\f\r\"\\");
328 const wxChar *escapes = wxT("abtnvfr\"\\");
329 wxString str;
330
331 for (size_t i = 0; i < arg.length(); i++) {
332 wxChar ch = (wxChar)arg[i];
333 const wxChar *p = wxStrchr(needEscape, ch);
334
335 if (p)
336 str += wxString::Format(wxT("\\%c"), escapes[p - needEscape]);
337 else if (wxIscntrl(ch))
338 str += wxString::Format(wxT("\\%03o"), ch);
339 else
340 str += (wxChar)ch;
341 }
342
343 return str.length() == arg.length() && str.find(' ') == wxString::npos ?
344 str : wxT("\"") + str + wxT("\"");
345 }
346
347
348 ///////////////////////////////////////////////////////////////////////////////
349 // Test suite
350
351 class RegExTestSuite : public TestSuite
352 {
353 public:
354 RegExTestSuite(string name) : TestSuite(name) { }
355 void add(const char *mode, const char *id, const char *flags,
356 const char *pattern, const char *data, const char *expected, ...);
357 };
358
359 // Add a testcase to the suite
360 //
361 void RegExTestSuite::add(
362 const char *mode,
363 const char *id,
364 const char *flags,
365 const char *pattern,
366 const char *data,
367 const char *expected, ...)
368 {
369 string name = getName() + "." + id;
370
371 vector<const char *> expected_results;
372 va_list ap;
373
374 for (va_start(ap, expected); expected; expected = va_arg(ap, const char *))
375 expected_results.push_back(expected);
376
377 va_end(ap);
378
379 try {
380 addTest(new RegExTestCase(
381 name, mode, id, flags, pattern, data, expected_results));
382 }
383 catch (Exception& e) {
384 wxLogInfo(wxString::Format(wxT("skipping: %s\n %s\n"),
385 wxString(name.c_str(), wxConvUTF8).c_str(),
386 wxString(e.what(), wxConvUTF8).c_str()));
387 }
388 }
389
390
391 // Include the generated tests
392 //
393 #include "regex.inc"
394
395
396 #endif // wxHAS_REGEX_ADVANCED
397
398 #endif // wxUSE_REGEX