1 | // ============================================================= // |
---|
2 | // // |
---|
3 | // File : RegExpr.cxx // |
---|
4 | // Purpose : Wrapper for ARBDB regular expressions // |
---|
5 | // // |
---|
6 | // Coded by Ralf Westram (coder@reallysoft.de) in April 2009 // |
---|
7 | // Institute of Microbiology (Technical University Munich) // |
---|
8 | // http://www.arb-home.de/ // |
---|
9 | // // |
---|
10 | // ============================================================= // |
---|
11 | |
---|
12 | #include "RegExpr.hxx" |
---|
13 | |
---|
14 | #include <arb_match.h> |
---|
15 | #include <arb_mem.h> |
---|
16 | |
---|
17 | #include <regex.h> |
---|
18 | |
---|
19 | using namespace std; |
---|
20 | |
---|
21 | struct GBS_regex { regex_t compiled; }; // definition exists twice (see ../../ARBDB/admatch.c) |
---|
22 | |
---|
23 | RegExpr::RegExpr(const std::string& expression_, bool ignore_case_) : |
---|
24 | expression(expression_), |
---|
25 | ignore_case(ignore_case_), |
---|
26 | comreg(NULp), |
---|
27 | matches(NULp), |
---|
28 | failure(NULp) |
---|
29 | {} |
---|
30 | |
---|
31 | RegExpr::~RegExpr() { |
---|
32 | if (comreg) GBS_free_regexpr(comreg); |
---|
33 | if (failure) delete failure; |
---|
34 | delete [] matches; |
---|
35 | } |
---|
36 | |
---|
37 | bool RegExpr::compile() const { |
---|
38 | if (!comreg && !failure) { |
---|
39 | delete [] matches; matches = NULp; |
---|
40 | |
---|
41 | GB_ERROR error = NULp; |
---|
42 | comreg = GBS_compile_regexpr(expression.c_str(), ignore_case ? GB_IGNORE_CASE : GB_MIND_CASE, &error); |
---|
43 | if (error) { |
---|
44 | failure = new string(error); |
---|
45 | } |
---|
46 | re_assert(contradicted(comreg, failure)); |
---|
47 | } |
---|
48 | return comreg; |
---|
49 | } |
---|
50 | |
---|
51 | void RegExpr::perform_match(const char *str, size_t offset) const { |
---|
52 | /* Searches for first match (and submatches) in 'str' |
---|
53 | * |
---|
54 | * sets member 'matches' to array of match + subexpression matches (heap-copy) |
---|
55 | * or to NULp if nothing matched |
---|
56 | * |
---|
57 | * If 'offset' > 0, then str is searched from position 'offset'. |
---|
58 | * In this case it is assumed, that we are not at line start! |
---|
59 | */ |
---|
60 | |
---|
61 | delete [] matches; matches = NULp; |
---|
62 | |
---|
63 | size_t subs = subexpr_count(); |
---|
64 | regmatch_t *possMatch = ARB_alloc<regmatch_t>(subs+1); |
---|
65 | int eflags = offset ? REG_NOTBOL : 0; |
---|
66 | int res = regexec(&comreg->compiled, str+offset, subs+1, possMatch, eflags); |
---|
67 | |
---|
68 | if (res != REG_NOMATCH) { |
---|
69 | matches = new RegMatch[subs+1]; |
---|
70 | for (size_t s = 0; s <= subs; s++) { |
---|
71 | if (possMatch[s].rm_so != -1) { // real match |
---|
72 | matches[s] = RegMatch(possMatch[s].rm_so+offset, possMatch[s].rm_eo+offset); |
---|
73 | } |
---|
74 | } |
---|
75 | re_assert(matches[0].didMatch()); // complete match has to be found |
---|
76 | } |
---|
77 | free(possMatch); |
---|
78 | } |
---|
79 | |
---|
80 | const RegMatch *RegExpr::match(const std::string& versus, size_t offset) const { |
---|
81 | if (!comreg) { |
---|
82 | // lazy compilation |
---|
83 | if (!compile()) return NULp; |
---|
84 | } |
---|
85 | perform_match(versus.c_str(), offset); |
---|
86 | return (matches && matches[0].didMatch()) ? &matches[0] : NULp; |
---|
87 | } |
---|
88 | |
---|
89 | size_t RegExpr::subexpr_count() const { |
---|
90 | if (!comreg) { |
---|
91 | // lazy compilation |
---|
92 | if (!compile()) return 0; |
---|
93 | } |
---|
94 | return comreg->compiled.re_nsub; |
---|
95 | } |
---|
96 | |
---|
97 | const RegMatch *RegExpr::subexpr_match(size_t subnr) const { |
---|
98 | // get subexpression match from last 'match()' |
---|
99 | // (or NULp if subexpression 'subnr' did not match) |
---|
100 | // |
---|
101 | // 'subnr' is in range [1..subexpr_count()] |
---|
102 | |
---|
103 | const RegMatch *result = NULp; |
---|
104 | if (matches) { |
---|
105 | size_t subs = subexpr_count(); |
---|
106 | re_assert(subnr >= 1 && subnr <= subs); // illegal subexpression index |
---|
107 | if (subnr >= 1 && subnr <= subs) { |
---|
108 | if (matches[subnr].didMatch()) result = &matches[subnr]; |
---|
109 | } |
---|
110 | } |
---|
111 | return result; |
---|
112 | } |
---|
113 | |
---|
114 | // -------------------------------------------------------------------------------- |
---|
115 | |
---|
116 | #ifdef UNIT_TESTS |
---|
117 | #ifndef TEST_UNIT_H |
---|
118 | #include <test_unit.h> |
---|
119 | #endif |
---|
120 | |
---|
121 | static arb_test::match_expectation got_a_match(const RegExpr& exp, const RegMatch *match) { |
---|
122 | using namespace arb_test; |
---|
123 | const string *expression_failed = exp.has_failed(); |
---|
124 | |
---|
125 | expectation_group expected(that(match).does_differ_from_NULL()); |
---|
126 | expected.add(that(expression_failed).is_equal_to_NULL()); |
---|
127 | if (expression_failed) { |
---|
128 | const char *expression_failed_msg = expression_failed->c_str(); |
---|
129 | expected.add(that(expression_failed_msg).is_equal_to("(unwanted)")); |
---|
130 | } |
---|
131 | return all().ofgroup(expected); |
---|
132 | } |
---|
133 | static arb_test::match_expectation got_no_match(const RegExpr& exp, const RegMatch *match) { |
---|
134 | using namespace arb_test; |
---|
135 | const string *expression_failed = exp.has_failed(); |
---|
136 | |
---|
137 | expectation_group expected(that(match).is_equal_to_NULL()); |
---|
138 | expected.add(that(expression_failed).is_equal_to_NULL()); |
---|
139 | if (expression_failed) { |
---|
140 | const char *expression_failed_msg = expression_failed->c_str(); |
---|
141 | expected.add(that(expression_failed_msg).is_equal_to("(unwanted)")); |
---|
142 | } |
---|
143 | return all().ofgroup(expected); |
---|
144 | } |
---|
145 | static arb_test::match_expectation fails_to_compile(const RegExpr& exp,const char *error) { |
---|
146 | using namespace arb_test; |
---|
147 | const string *expression_failed = exp.has_failed(); |
---|
148 | |
---|
149 | expectation_group expected(that(expression_failed).does_differ_from_NULL()); |
---|
150 | if (expression_failed) { |
---|
151 | const char *expression_failed_msg = expression_failed->c_str(); |
---|
152 | expected.add(that(expression_failed_msg).is_equal_to(error)); |
---|
153 | } |
---|
154 | return all().ofgroup(expected); |
---|
155 | } |
---|
156 | |
---|
157 | #define TEST_REGEX_FAILS_TO_COMPILE(regexpr,error) do{ \ |
---|
158 | RegExpr exp(regexpr, false); \ |
---|
159 | TEST_EXPECTATION(fails_to_compile(exp,error)); \ |
---|
160 | }while(0) |
---|
161 | |
---|
162 | #define TEST_REGEX_MATCHES(str,regexpr,igCase,exp_match) do { \ |
---|
163 | RegExpr exp(regexpr, igCase); \ |
---|
164 | const RegMatch *match = exp.match(str); \ |
---|
165 | TEST_EXPECTATION(got_a_match(exp,match)); \ |
---|
166 | TEST_EXPECT_EQUAL(match->extract(str).c_str(), exp_match); \ |
---|
167 | } while(0) |
---|
168 | |
---|
169 | #define TEST_REGEX_DOESNT_MATCH(str,regexpr,igCase) do { \ |
---|
170 | RegExpr exp(regexpr, igCase); \ |
---|
171 | const RegMatch *match = exp.match(str); \ |
---|
172 | TEST_EXPECTATION(got_no_match(exp,match)); \ |
---|
173 | } while(0) |
---|
174 | |
---|
175 | #define TEST_REGEX_MATCHES_SUB1(str,regexpr,igCase,exp_match,exp_sub1match) do { \ |
---|
176 | RegExpr exp(regexpr, igCase); \ |
---|
177 | const RegMatch *match = exp.match(str); \ |
---|
178 | TEST_EXPECTATION(got_a_match(exp,match)); \ |
---|
179 | TEST_EXPECT_EQUAL(match->extract(str).c_str(), exp_match); \ |
---|
180 | match = exp.subexpr_match(1); \ |
---|
181 | TEST_REJECT_NULL(match); \ |
---|
182 | TEST_EXPECT_EQUAL(match->extract(str).c_str(), exp_sub1match); \ |
---|
183 | } while(0) |
---|
184 | |
---|
185 | void TEST_regexpr() { |
---|
186 | TEST_REGEX_MATCHES("bla", "^bla$", false, "bla"); |
---|
187 | |
---|
188 | TEST_REGEX_DOESNT_MATCH("3;1406", "^bla$", true); |
---|
189 | |
---|
190 | // disable the following section at will (arb does not depend on it) |
---|
191 | #if defined(DARWIN) |
---|
192 | TEST_REGEX_FAILS_TO_COMPILE("*", "repetition-operator operand invalid"); |
---|
193 | TEST_REGEX_FAILS_TO_COMPILE("[[:fantasy:]]*", "invalid character class"); |
---|
194 | TEST_REGEX_FAILS_TO_COMPILE("x{2", "braces not balanced"); |
---|
195 | TEST_REGEX_FAILS_TO_COMPILE("x{2-5}", "invalid repetition count(s)"); |
---|
196 | TEST_REGEX_FAILS_TO_COMPILE("x{5,2}", "invalid repetition count(s)"); |
---|
197 | #else // !DARWIN |
---|
198 | TEST_REGEX_FAILS_TO_COMPILE("*", "Invalid preceding regular expression"); |
---|
199 | TEST_REGEX_FAILS_TO_COMPILE("[[:fantasy:]]*", "Invalid character class name"); |
---|
200 | TEST_REGEX_FAILS_TO_COMPILE("x{2", "Unmatched \\{"); |
---|
201 | TEST_REGEX_FAILS_TO_COMPILE("x{2-5}", "Invalid content of \\{\\}"); |
---|
202 | TEST_REGEX_FAILS_TO_COMPILE("x{5,2}", "Invalid content of \\{\\}"); |
---|
203 | #endif |
---|
204 | |
---|
205 | // RESULT_MODIFIED_OSX: regex isn't portable :/ |
---|
206 | #if defined(DARWIN) |
---|
207 | TEST_REGEX_FAILS_TO_COMPILE("^bla|", "empty (sub)expression"); |
---|
208 | #else // !DARWIN |
---|
209 | TEST_REGEX_MATCHES("3;1406", "^bla|", true, ""); |
---|
210 | #endif |
---|
211 | |
---|
212 | TEST_REGEX_MATCHES_SUB1("3;1406", "^[0-9]+;([0-9]+)$", true, "3;1406", "1406"); |
---|
213 | TEST_REGEX_MATCHES_SUB1(" Find CAPITAL \t WoRD ", "[[:space:]]+([A-Z]+)[[:space:]]+", false, " CAPITAL \t ", "CAPITAL"); |
---|
214 | TEST_REGEX_MATCHES_SUB1("--ajsd--aksjdh----alsdjkasldja---", "-+([a-z]{5,8})-+", true, "--aksjdh----", "aksjdh"); |
---|
215 | } |
---|
216 | TEST_PUBLISH(TEST_regexpr); |
---|
217 | |
---|
218 | #endif // UNIT_TESTS |
---|
219 | |
---|
220 | // -------------------------------------------------------------------------------- |
---|