1*b0d29bc4SBrooks Davis // Copyright 2014 The Kyua Authors.
2*b0d29bc4SBrooks Davis // All rights reserved.
3*b0d29bc4SBrooks Davis //
4*b0d29bc4SBrooks Davis // Redistribution and use in source and binary forms, with or without
5*b0d29bc4SBrooks Davis // modification, are permitted provided that the following conditions are
6*b0d29bc4SBrooks Davis // met:
7*b0d29bc4SBrooks Davis //
8*b0d29bc4SBrooks Davis // * Redistributions of source code must retain the above copyright
9*b0d29bc4SBrooks Davis // notice, this list of conditions and the following disclaimer.
10*b0d29bc4SBrooks Davis // * Redistributions in binary form must reproduce the above copyright
11*b0d29bc4SBrooks Davis // notice, this list of conditions and the following disclaimer in the
12*b0d29bc4SBrooks Davis // documentation and/or other materials provided with the distribution.
13*b0d29bc4SBrooks Davis // * Neither the name of Google Inc. nor the names of its contributors
14*b0d29bc4SBrooks Davis // may be used to endorse or promote products derived from this software
15*b0d29bc4SBrooks Davis // without specific prior written permission.
16*b0d29bc4SBrooks Davis //
17*b0d29bc4SBrooks Davis // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18*b0d29bc4SBrooks Davis // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19*b0d29bc4SBrooks Davis // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20*b0d29bc4SBrooks Davis // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21*b0d29bc4SBrooks Davis // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22*b0d29bc4SBrooks Davis // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23*b0d29bc4SBrooks Davis // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24*b0d29bc4SBrooks Davis // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25*b0d29bc4SBrooks Davis // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26*b0d29bc4SBrooks Davis // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27*b0d29bc4SBrooks Davis // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28*b0d29bc4SBrooks Davis
29*b0d29bc4SBrooks Davis #include "utils/text/regex.hpp"
30*b0d29bc4SBrooks Davis
31*b0d29bc4SBrooks Davis #include <atf-c++.hpp>
32*b0d29bc4SBrooks Davis
33*b0d29bc4SBrooks Davis #include "utils/text/exceptions.hpp"
34*b0d29bc4SBrooks Davis
35*b0d29bc4SBrooks Davis namespace text = utils::text;
36*b0d29bc4SBrooks Davis
37*b0d29bc4SBrooks Davis
38*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__no_matches);
ATF_TEST_CASE_BODY(integration__no_matches)39*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__no_matches)
40*b0d29bc4SBrooks Davis {
41*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
42*b0d29bc4SBrooks Davis "foo.*bar", "this is a string without the searched text", 0);
43*b0d29bc4SBrooks Davis ATF_REQUIRE(!matches);
44*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(0, matches.count());
45*b0d29bc4SBrooks Davis }
46*b0d29bc4SBrooks Davis
47*b0d29bc4SBrooks Davis
48*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__no_capture_groups);
ATF_TEST_CASE_BODY(integration__no_capture_groups)49*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__no_capture_groups)
50*b0d29bc4SBrooks Davis {
51*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
52*b0d29bc4SBrooks Davis "foo.*bar", "this is a string with foo and bar embedded in it", 0);
53*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
54*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(1, matches.count());
55*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("foo and bar", matches.get(0));
56*b0d29bc4SBrooks Davis }
57*b0d29bc4SBrooks Davis
58*b0d29bc4SBrooks Davis
59*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__one_capture_group);
ATF_TEST_CASE_BODY(integration__one_capture_group)60*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__one_capture_group)
61*b0d29bc4SBrooks Davis {
62*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
63*b0d29bc4SBrooks Davis "^([^ ]*) ", "the string", 1);
64*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
65*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(2, matches.count());
66*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("the ", matches.get(0));
67*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("the", matches.get(1));
68*b0d29bc4SBrooks Davis }
69*b0d29bc4SBrooks Davis
70*b0d29bc4SBrooks Davis
71*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__many_capture_groups);
ATF_TEST_CASE_BODY(integration__many_capture_groups)72*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__many_capture_groups)
73*b0d29bc4SBrooks Davis {
74*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
75*b0d29bc4SBrooks Davis "is ([^ ]*) ([a-z]*) to", "this is another string to parse", 2);
76*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
77*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(3, matches.count());
78*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("is another string to", matches.get(0));
79*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("another", matches.get(1));
80*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("string", matches.get(2));
81*b0d29bc4SBrooks Davis }
82*b0d29bc4SBrooks Davis
83*b0d29bc4SBrooks Davis
84*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__capture_groups_underspecified);
ATF_TEST_CASE_BODY(integration__capture_groups_underspecified)85*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__capture_groups_underspecified)
86*b0d29bc4SBrooks Davis {
87*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
88*b0d29bc4SBrooks Davis "is ([^ ]*) ([a-z]*) to", "this is another string to parse", 1);
89*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
90*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(2, matches.count());
91*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("is another string to", matches.get(0));
92*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("another", matches.get(1));
93*b0d29bc4SBrooks Davis }
94*b0d29bc4SBrooks Davis
95*b0d29bc4SBrooks Davis
96*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__capture_groups_overspecified);
ATF_TEST_CASE_BODY(integration__capture_groups_overspecified)97*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__capture_groups_overspecified)
98*b0d29bc4SBrooks Davis {
99*b0d29bc4SBrooks Davis const text::regex_matches matches = text::match_regex(
100*b0d29bc4SBrooks Davis "is ([^ ]*) ([a-z]*) to", "this is another string to parse", 10);
101*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
102*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(3, matches.count());
103*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("is another string to", matches.get(0));
104*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("another", matches.get(1));
105*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("string", matches.get(2));
106*b0d29bc4SBrooks Davis }
107*b0d29bc4SBrooks Davis
108*b0d29bc4SBrooks Davis
109*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__reuse_regex_in_multiple_matches);
ATF_TEST_CASE_BODY(integration__reuse_regex_in_multiple_matches)110*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__reuse_regex_in_multiple_matches)
111*b0d29bc4SBrooks Davis {
112*b0d29bc4SBrooks Davis const text::regex regex = text::regex::compile("number is ([0-9]+)", 1);
113*b0d29bc4SBrooks Davis
114*b0d29bc4SBrooks Davis {
115*b0d29bc4SBrooks Davis const text::regex_matches matches = regex.match("my number is 581.");
116*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
117*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(2, matches.count());
118*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("number is 581", matches.get(0));
119*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("581", matches.get(1));
120*b0d29bc4SBrooks Davis }
121*b0d29bc4SBrooks Davis
122*b0d29bc4SBrooks Davis {
123*b0d29bc4SBrooks Davis const text::regex_matches matches = regex.match("your number is 6");
124*b0d29bc4SBrooks Davis ATF_REQUIRE(matches);
125*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ(2, matches.count());
126*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("number is 6", matches.get(0));
127*b0d29bc4SBrooks Davis ATF_REQUIRE_EQ("6", matches.get(1));
128*b0d29bc4SBrooks Davis }
129*b0d29bc4SBrooks Davis }
130*b0d29bc4SBrooks Davis
131*b0d29bc4SBrooks Davis
132*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__ignore_case);
ATF_TEST_CASE_BODY(integration__ignore_case)133*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__ignore_case)
134*b0d29bc4SBrooks Davis {
135*b0d29bc4SBrooks Davis const text::regex regex1 = text::regex::compile("foo", 0, false);
136*b0d29bc4SBrooks Davis ATF_REQUIRE(!regex1.match("bar Foo bar"));
137*b0d29bc4SBrooks Davis ATF_REQUIRE(!regex1.match("bar foO bar"));
138*b0d29bc4SBrooks Davis ATF_REQUIRE(!regex1.match("bar FOO bar"));
139*b0d29bc4SBrooks Davis
140*b0d29bc4SBrooks Davis ATF_REQUIRE(!text::match_regex("foo", "bar Foo bar", 0, false));
141*b0d29bc4SBrooks Davis ATF_REQUIRE(!text::match_regex("foo", "bar foO bar", 0, false));
142*b0d29bc4SBrooks Davis ATF_REQUIRE(!text::match_regex("foo", "bar FOO bar", 0, false));
143*b0d29bc4SBrooks Davis
144*b0d29bc4SBrooks Davis const text::regex regex2 = text::regex::compile("foo", 0, true);
145*b0d29bc4SBrooks Davis ATF_REQUIRE( regex2.match("bar foo bar"));
146*b0d29bc4SBrooks Davis ATF_REQUIRE( regex2.match("bar Foo bar"));
147*b0d29bc4SBrooks Davis ATF_REQUIRE( regex2.match("bar foO bar"));
148*b0d29bc4SBrooks Davis ATF_REQUIRE( regex2.match("bar FOO bar"));
149*b0d29bc4SBrooks Davis
150*b0d29bc4SBrooks Davis ATF_REQUIRE( text::match_regex("foo", "bar foo bar", 0, true));
151*b0d29bc4SBrooks Davis ATF_REQUIRE( text::match_regex("foo", "bar Foo bar", 0, true));
152*b0d29bc4SBrooks Davis ATF_REQUIRE( text::match_regex("foo", "bar foO bar", 0, true));
153*b0d29bc4SBrooks Davis ATF_REQUIRE( text::match_regex("foo", "bar FOO bar", 0, true));
154*b0d29bc4SBrooks Davis }
155*b0d29bc4SBrooks Davis
156*b0d29bc4SBrooks Davis ATF_TEST_CASE_WITHOUT_HEAD(integration__invalid_regex);
ATF_TEST_CASE_BODY(integration__invalid_regex)157*b0d29bc4SBrooks Davis ATF_TEST_CASE_BODY(integration__invalid_regex)
158*b0d29bc4SBrooks Davis {
159*b0d29bc4SBrooks Davis ATF_REQUIRE_THROW(text::regex_error,
160*b0d29bc4SBrooks Davis text::regex::compile("this is (unbalanced", 0));
161*b0d29bc4SBrooks Davis }
162*b0d29bc4SBrooks Davis
163*b0d29bc4SBrooks Davis
ATF_INIT_TEST_CASES(tcs)164*b0d29bc4SBrooks Davis ATF_INIT_TEST_CASES(tcs)
165*b0d29bc4SBrooks Davis {
166*b0d29bc4SBrooks Davis // regex and regex_matches are so coupled that it makes no sense to test
167*b0d29bc4SBrooks Davis // them independently. Just validate their integration.
168*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__no_matches);
169*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__no_capture_groups);
170*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__one_capture_group);
171*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__many_capture_groups);
172*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__capture_groups_underspecified);
173*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__capture_groups_overspecified);
174*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__reuse_regex_in_multiple_matches);
175*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__ignore_case);
176*b0d29bc4SBrooks Davis ATF_ADD_TEST_CASE(tcs, integration__invalid_regex);
177*b0d29bc4SBrooks Davis }
178