1 // Copyright 2015 The PDFium Authors
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
4
5 #include "core/fpdftext/cpdf_linkextract.h"
6
7 #include <utility>
8
9 #include "testing/gtest/include/gtest/gtest.h"
10
11 // Class to help test functions in CPDF_LinkExtract class.
12 class CPDF_TestLinkExtract final : public CPDF_LinkExtract {
13 public:
CPDF_TestLinkExtract()14 CPDF_TestLinkExtract() : CPDF_LinkExtract(nullptr) {}
15
16 private:
17 // Add test cases as friends to access protected member functions.
18 // Access CheckMailLink and CheckWebLink.
19 FRIEND_TEST(CPDFLinkExtractTest, CheckMailLink);
20 FRIEND_TEST(CPDFLinkExtractTest, CheckWebLink);
21 };
22
TEST(CPDFLinkExtractTest,CheckMailLink)23 TEST(CPDFLinkExtractTest, CheckMailLink) {
24 CPDF_TestLinkExtract extractor;
25 // Check cases that fail to extract valid mail link.
26 const wchar_t* const kInvalidStrings[] = {
27 L"",
28 L"peter.pan", // '@' is required.
29 L"abc@server", // Domain name needs at least one '.'.
30 L"abc.@gmail.com", // '.' can not immediately precede '@'.
31 L"abc@xyz&q.org", // Domain name should not contain '&'.
32 L"abc@.xyz.org", // Domain name should not start with '.'.
33 L"fan@g..com" // Domain name should not have consecutive '.'
34 };
35 for (const wchar_t* input : kInvalidStrings) {
36 WideString text_str(input);
37 EXPECT_FALSE(extractor.CheckMailLink(&text_str)) << input;
38 }
39
40 // A struct of {input_string, expected_extracted_email_address}.
41 struct IOPair {
42 const wchar_t* input;
43 const wchar_t* expected_output;
44 };
45 // Check cases that can extract valid mail link.
46 constexpr IOPair kValidStrings[] = {
47 {L"peter@abc.d", L"peter@abc.d"},
48 {L"red.teddy.b@abc.com", L"red.teddy.b@abc.com"},
49 {L"abc_@gmail.com", L"abc_@gmail.com"}, // '_' is ok before '@'.
50 {L"dummy-hi@gmail.com",
51 L"dummy-hi@gmail.com"}, // '-' is ok in user name.
52 {L"a..df@gmail.com", L"df@gmail.com"}, // Stop at consecutive '.'.
53 {L".john@yahoo.com", L"john@yahoo.com"}, // Remove heading '.'.
54 {L"abc@xyz.org?/", L"abc@xyz.org"}, // Trim ending invalid chars.
55 {L"fan{abc@xyz.org", L"abc@xyz.org"}, // Trim beginning invalid chars.
56 {L"fan@g.com..", L"fan@g.com"}, // Trim the ending periods.
57 {L"CAP.cap@Gmail.Com", L"CAP.cap@Gmail.Com"}, // Keep the original case.
58 };
59 for (const auto& it : kValidStrings) {
60 WideString text_str(it.input);
61 WideString expected_str(L"mailto:");
62 expected_str += it.expected_output;
63 EXPECT_TRUE(extractor.CheckMailLink(&text_str)) << it.input;
64 EXPECT_EQ(expected_str.c_str(), text_str);
65 }
66 }
67
TEST(CPDFLinkExtractTest,CheckWebLink)68 TEST(CPDFLinkExtractTest, CheckWebLink) {
69 CPDF_TestLinkExtract extractor;
70 // Check cases that fail to extract valid web link.
71 // The last few are legit web addresses that we don't handle now.
72 const wchar_t* const kInvalidCases[] = {
73 L"", // Blank.
74 L"http", // No colon.
75 L"www.", // Missing domain.
76 L"https-and-www", // Dash not colon.
77 L"http:/abc.com", // Missing slash.
78 L"http://((()),", // Only invalid chars in host name.
79 L"ftp://example.com", // Ftp scheme is not supported.
80 L"http:example.com", // Missing slashes.
81 L"http//[example.com", // Invalid IPv6 address.
82 L"http//[00:00:00:00:00:00", // Invalid IPv6 address.
83 L"http//[]", // Empty IPv6 address.
84 L"abc.example.com", // URL without scheme.
85 };
86 for (const wchar_t* input : kInvalidCases) {
87 auto maybe_link = extractor.CheckWebLink(input);
88 EXPECT_FALSE(maybe_link.has_value()) << input;
89 }
90
91 // Check cases that can extract valid web link.
92 // An array of {input_string, expected_extracted_web_link}.
93 struct ValidCase {
94 const wchar_t* const input_string;
95 const wchar_t* const url_extracted;
96 const size_t start_offset;
97 const size_t count;
98 };
99 const ValidCase kValidCases[] = {
100 {L"http://www.example.com", L"http://www.example.com", 0,
101 22}, // standard URL.
102 {L"http://www.example.com:88", L"http://www.example.com:88", 0,
103 25}, // URL with port number.
104 {L"http://test@www.example.com", L"http://test@www.example.com", 0,
105 27}, // URL with username.
106 {L"http://test:test@example.com", L"http://test:test@example.com", 0,
107 28}, // URL with username and password.
108 {L"http://example", L"http://example", 0,
109 14}, // URL with short domain name.
110 {L"http////www.server", L"http://www.server", 8,
111 10}, // URL starts with "www.".
112 {L"http:/www.abc.com", L"http://www.abc.com", 6,
113 11}, // URL starts with "www.".
114 {L"www.a.b.c", L"http://www.a.b.c", 0, 9}, // URL starts with "www.".
115 {L"https://a.us", L"https://a.us", 0, 12}, // Secure http URL.
116 {L"https://www.t.us", L"https://www.t.us", 0, 16}, // Secure http URL.
117 {L"www.example-test.com", L"http://www.example-test.com", 0,
118 20}, // '-' in host is ok.
119 {L"www.example.com,", L"http://www.example.com", 0,
120 15}, // Trim ending invalid chars.
121 {L"www.example.com;(", L"http://www.example.com", 0,
122 15}, // Trim ending invalid chars.
123 {L"test:www.abc.com", L"http://www.abc.com", 5,
124 11}, // Trim chars before URL.
125 {L"(http://www.abc.com)", L"http://www.abc.com", 1,
126 18}, // Trim external brackets.
127 {L"0(http://www.abc.com)0", L"http://www.abc.com", 2,
128 18}, // Trim chars outside brackets as well.
129 {L"0(www.abc.com)0", L"http://www.abc.com", 2,
130 11}, // Links without http should also have brackets trimmed.
131 {L"http://www.abc.com)0", L"http://www.abc.com)0", 0,
132 20}, // Do not trim brackets that were not opened.
133 {L"{(<http://www.abc.com>)}", L"http://www.abc.com", 3,
134 18}, // Trim chars with multiple levels of brackets.
135 {L"[http://www.abc.com/z(1)]", L"http://www.abc.com/z(1)", 1,
136 23}, // Brackets opened inside the URL should not be trimmed.
137 {L"(http://www.abc.com/z(1))", L"http://www.abc.com/z(1)", 1,
138 23}, // Brackets opened inside the URL should not be trimmed.
139 {L"\"http://www.abc.com\"", L"http://www.abc.com", 1,
140 18}, // External quotes can also be escaped
141 {L"www.g.com..", L"http://www.g.com..", 0, 11}, // Leave ending periods.
142
143 // Web links can contain IP addresses too.
144 {L"http://192.168.0.1", L"http://192.168.0.1", 0, 18}, // IPv4 address.
145 {L"http://192.168.0.1:80", L"http://192.168.0.1:80", 0,
146 21}, // IPv4 address with port.
147 {L"http://[aa::00:bb::00:cc:00]", L"http://[aa::00:bb::00:cc:00]", 0,
148 28}, // IPv6 reference.
149 {L"http://[aa::00:bb::00:cc:00]:12", L"http://[aa::00:bb::00:cc:00]:12",
150 0, 31}, // IPv6 reference with port.
151 {L"http://[aa]:12", L"http://[aa]:12", 0,
152 14}, // Not validate IP address.
153 {L"http://[aa]:12abc", L"http://[aa]:12", 0,
154 14}, // Trim for IPv6 address.
155 {L"http://[aa]:", L"http://[aa]", 0, 11}, // Trim for IPv6 address.
156
157 // Path and query parts can be anything.
158 {L"www.abc.com/#%%^&&*(", L"http://www.abc.com/#%%^&&*(", 0, 20},
159 {L"www.a.com/#a=@?q=rr&r=y", L"http://www.a.com/#a=@?q=rr&r=y", 0, 23},
160 {L"http://a.com/1/2/3/4\5\6", L"http://a.com/1/2/3/4\5\6", 0, 22},
161 {L"http://www.example.com/foo;bar", L"http://www.example.com/foo;bar", 0,
162 30},
163
164 // Invalid chars inside host name are ok as we don't validate them.
165 {L"http://ex[am]ple", L"http://ex[am]ple", 0, 16},
166 {L"http://:example.com", L"http://:example.com", 0, 19},
167 {L"http://((())/path?", L"http://((())/path?", 0, 18},
168 {L"http:////abc.server", L"http:////abc.server", 0, 19},
169
170 // Non-ASCII chars are not validated either.
171 {L"www.测试.net", L"http://www.测试.net", 0, 10},
172 {L"www.测试。net。", L"http://www.测试。net。", 0, 11},
173 {L"www.测试.net；", L"http://www.测试.net；", 0, 11},
174 };
175 for (const auto& it : kValidCases) {
176 auto maybe_link = extractor.CheckWebLink(it.input_string);
177 ASSERT_TRUE(maybe_link.has_value()) << it.input_string;
178 EXPECT_EQ(it.url_extracted, maybe_link.value().m_strUrl);
179 EXPECT_EQ(it.start_offset, maybe_link.value().m_Start) << it.input_string;
180 EXPECT_EQ(it.count, maybe_link.value().m_Count) << it.input_string;
181 }
182 }
183