1 // Copyright 2013 The Chromium 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 <stddef.h>
6
7 #include "testing/gtest/include/gtest/gtest.h"
8 #include "url/third_party/mozilla/url_parse.h"
9
10 // Interesting IE file:isms...
11 //
12 // file:/foo/bar file:///foo/bar
13 // The result here seems totally invalid!?!? This isn't UNC.
14 //
15 // file:/
16 // file:// or any other number of slashes
17 // IE6 doesn't do anything at all if you click on this link. No error:
18 // nothing. IE6's history system seems to always color this link, so I'm
19 // guessing that it maps internally to the empty URL.
20 //
21 // C:\ file:///C:/
22 // / file:///C:/
23 // /foo file:///C:/foo
24 // Interestingly, IE treats "/" as an alias for "c:\", which makes sense,
25 // but is weird to think about on Windows.
26 //
27 // file:foo/ file:foo/ (invalid?!?!?)
28 // file:/foo/ file:///foo/ (invalid?!?!?)
29 // file://foo/ file://foo/ (UNC to server "foo")
30 // file:///foo/ file:///foo/ (invalid)
31 // file:////foo/ file://foo/ (UNC to server "foo")
32 // Any more than four slashes is also treated as UNC.
33 //
34 // file:C:/ file://C:/
35 // file:/C:/ file://C:/
36 // The number of slashes after "file:" don't matter if the thing following
37 // it looks like an absolute drive path. Also, slashes and backslashes are
38 // equally valid here.
39
40 namespace url {
41
42 namespace {
43
44 // Used for regular URL parse cases.
45 struct URLParseCase {
46 const char* input;
47
48 const char* scheme;
49 const char* username;
50 const char* password;
51 const char* host;
52 int port;
53 const char* path;
54 const char* query;
55 const char* ref;
56 };
57
58 // Simpler version of URLParseCase for testing path URLs.
59 struct PathURLParseCase {
60 const char* input;
61
62 const char* scheme;
63 const char* path;
64 };
65
66 // Simpler version of URLParseCase for testing mailto URLs.
67 struct MailtoURLParseCase {
68 const char* input;
69
70 const char* scheme;
71 const char* path;
72 const char* query;
73 };
74
75 // More complicated version of URLParseCase for testing filesystem URLs.
76 struct FileSystemURLParseCase {
77 const char* input;
78
79 const char* inner_scheme;
80 const char* inner_username;
81 const char* inner_password;
82 const char* inner_host;
83 int inner_port;
84 const char* inner_path;
85 const char* path;
86 const char* query;
87 const char* ref;
88 };
89
ComponentMatches(const char * input,const char * reference,const Component & component)90 bool ComponentMatches(const char* input,
91 const char* reference,
92 const Component& component) {
93 // Check that the -1 sentinel is the only allowed negative value.
94 EXPECT_TRUE(component.is_valid() || component.len == -1);
95
96 // Begin should be valid.
97 EXPECT_LE(0, component.begin);
98
99 // A NULL reference means the component should be nonexistent.
100 if (!reference)
101 return component.len == -1;
102 if (!component.is_valid())
103 return false; // Reference is not NULL but we don't have anything
104
105 if (strlen(reference) != static_cast<size_t>(component.len))
106 return false; // Lengths don't match
107
108 // Now check the actual characters.
109 return strncmp(reference, &input[component.begin], component.len) == 0;
110 }
111
ExpectInvalidComponent(const Component & component)112 void ExpectInvalidComponent(const Component& component) {
113 EXPECT_EQ(0, component.begin);
114 EXPECT_EQ(-1, component.len);
115 }
116
URLParseCaseMatches(const URLParseCase & expected,const Parsed & parsed)117 void URLParseCaseMatches(const URLParseCase& expected, const Parsed& parsed) {
118 const char* url = expected.input;
119 SCOPED_TRACE(testing::Message()
120 << "url: \"" << url << "\", parsed: " << parsed);
121 int port = ParsePort(url, parsed.port);
122 EXPECT_TRUE(ComponentMatches(url, expected.scheme, parsed.scheme));
123 EXPECT_TRUE(ComponentMatches(url, expected.username, parsed.username));
124 EXPECT_TRUE(ComponentMatches(url, expected.password, parsed.password));
125 EXPECT_TRUE(ComponentMatches(url, expected.host, parsed.host));
126 EXPECT_EQ(expected.port, port);
127 EXPECT_TRUE(ComponentMatches(url, expected.path, parsed.path));
128 EXPECT_TRUE(ComponentMatches(url, expected.query, parsed.query));
129 EXPECT_TRUE(ComponentMatches(url, expected.ref, parsed.ref));
130 }
131
132 // Parsed ----------------------------------------------------------------------
133
TEST(URLParser,Length)134 TEST(URLParser, Length) {
135 const char* length_cases[] = {
136 // One with everything in it.
137 "http://user:pass@host:99/foo?bar#baz",
138 // One with nothing in it.
139 "",
140 // Working backwards, let's start taking off stuff from the full one.
141 "http://user:pass@host:99/foo?bar#",
142 "http://user:pass@host:99/foo?bar",
143 "http://user:pass@host:99/foo?",
144 "http://user:pass@host:99/foo",
145 "http://user:pass@host:99/",
146 "http://user:pass@host:99",
147 "http://user:pass@host:",
148 "http://user:pass@host",
149 "http://host",
150 "http://user@",
151 "http:",
152 };
153 for (const char* length_case : length_cases) {
154 int true_length = static_cast<int>(strlen(length_case));
155
156 Parsed parsed;
157 ParseStandardURL(length_case, true_length, &parsed);
158
159 EXPECT_EQ(true_length, parsed.Length());
160 }
161 }
162
TEST(URLParser,CountCharactersBefore)163 TEST(URLParser, CountCharactersBefore) {
164 struct CountCase {
165 const char* url;
166 Parsed::ComponentType component;
167 bool include_delimiter;
168 int expected_count;
169 } count_cases[] = {
170 // Test each possibility in the case where all components are present.
171 // 0 1 2
172 // 0123456789012345678901
173 {"http://u:p@h:8/p?q#r", Parsed::SCHEME, true, 0},
174 {"http://u:p@h:8/p?q#r", Parsed::SCHEME, false, 0},
175 {"http://u:p@h:8/p?q#r", Parsed::USERNAME, true, 7},
176 {"http://u:p@h:8/p?q#r", Parsed::USERNAME, false, 7},
177 {"http://u:p@h:8/p?q#r", Parsed::PASSWORD, true, 9},
178 {"http://u:p@h:8/p?q#r", Parsed::PASSWORD, false, 9},
179 {"http://u:p@h:8/p?q#r", Parsed::HOST, true, 11},
180 {"http://u:p@h:8/p?q#r", Parsed::HOST, false, 11},
181 {"http://u:p@h:8/p?q#r", Parsed::PORT, true, 12},
182 {"http://u:p@h:8/p?q#r", Parsed::PORT, false, 13},
183 {"http://u:p@h:8/p?q#r", Parsed::PATH, false, 14},
184 {"http://u:p@h:8/p?q#r", Parsed::PATH, true, 14},
185 {"http://u:p@h:8/p?q#r", Parsed::QUERY, true, 16},
186 {"http://u:p@h:8/p?q#r", Parsed::QUERY, false, 17},
187 {"http://u:p@h:8/p?q#r", Parsed::REF, true, 18},
188 {"http://u:p@h:8/p?q#r", Parsed::REF, false, 19},
189 // Now test when the requested component is missing.
190 {"http://u:p@h:8/p?", Parsed::REF, true, 17},
191 {"http://u:p@h:8/p?q", Parsed::REF, true, 18},
192 {"http://u:p@h:8/p#r", Parsed::QUERY, true, 16},
193 {"http://u:p@h:8#r", Parsed::PATH, true, 14},
194 {"http://u:p@h/", Parsed::PORT, true, 12},
195 {"http://u:p@/", Parsed::HOST, true, 11},
196 // This case is a little weird. It will report that the password would
197 // start where the host begins. This is arguably correct, although you
198 // could also argue that it should start at the '@' sign. Doing it
199 // starting with the '@' sign is actually harder, so we don't bother.
200 {"http://u@h/", Parsed::PASSWORD, true, 9},
201 {"http://h/", Parsed::USERNAME, true, 7},
202 {"http:", Parsed::USERNAME, true, 5},
203 {"", Parsed::SCHEME, true, 0},
204 // Make sure a random component still works when there's nothing there.
205 {"", Parsed::REF, true, 0},
206 // File URLs are special with no host, so we test those.
207 {"file:///c:/foo", Parsed::USERNAME, true, 7},
208 {"file:///c:/foo", Parsed::PASSWORD, true, 7},
209 {"file:///c:/foo", Parsed::HOST, true, 7},
210 {"file:///c:/foo", Parsed::PATH, true, 7},
211 };
212 for (const auto& count_case : count_cases) {
213 int length = static_cast<int>(strlen(count_case.url));
214
215 // Simple test to distinguish file and standard URLs.
216 Parsed parsed;
217 if (length > 0 && count_case.url[0] == 'f') {
218 ParseFileURL(count_case.url, length, &parsed);
219 } else {
220 ParseStandardURL(count_case.url, length, &parsed);
221 }
222
223 int chars_before = parsed.CountCharactersBefore(
224 count_case.component, count_case.include_delimiter);
225 EXPECT_EQ(count_case.expected_count, chars_before);
226 }
227 }
228
229 // Standard --------------------------------------------------------------------
230
231 // clang-format off
232 // Input Scheme Usrname Passwd Host Port Path Query Ref
233 // ------------------------------------ ------- -------- ---------- ------------ --- ---------- ------------ -----
234 static URLParseCase cases[] = {
235 // Regular URL with all the parts
236 {"http://user:pass@foo:21/bar;par?b#c", "http", "user", "pass", "foo", 21, "/bar;par","b", "c"},
237
238 // Known schemes should lean towards authority identification
239 {"http:foo.com", "http", nullptr, nullptr, "foo.com", -1, nullptr, nullptr, nullptr},
240
241 // Spaces!
242 {"\t :foo.com \n", "", nullptr, nullptr, "foo.com", -1, nullptr, nullptr, nullptr},
243 {" foo.com ", nullptr,nullptr, nullptr, "foo.com", -1, nullptr, nullptr, nullptr},
244 {"a:\t foo.com", "a", nullptr, nullptr, "\t foo.com", -1, nullptr, nullptr, nullptr},
245 {"http://f:21/ b ? d # e ", "http", nullptr, nullptr, "f", 21, "/ b ", " d ", " e"},
246
247 // Invalid port numbers should be identified and turned into -2, empty port
248 // numbers should be -1. Spaces aren't allowed in port numbers
249 {"http://f:/c", "http", nullptr, nullptr, "f", -1, "/c", nullptr, nullptr},
250 {"http://f:0/c", "http", nullptr, nullptr, "f", 0, "/c", nullptr, nullptr},
251 {"http://f:00000000000000/c", "http", nullptr, nullptr, "f", 0, "/c", nullptr, nullptr},
252 {"http://f:00000000000000000000080/c", "http", nullptr, nullptr, "f", 80, "/c", nullptr, nullptr},
253 {"http://f:b/c", "http", nullptr, nullptr, "f", -2, "/c", nullptr, nullptr},
254 {"http://f: /c", "http", nullptr, nullptr, "f", -2, "/c", nullptr, nullptr},
255 {"http://f:\n/c", "http", nullptr, nullptr, "f", -2, "/c", nullptr, nullptr},
256 {"http://f:fifty-two/c", "http", nullptr, nullptr, "f", -2, "/c", nullptr, nullptr},
257 {"http://f:999999/c", "http", nullptr, nullptr, "f", -2, "/c", nullptr, nullptr},
258 {"http://f: 21 / b ? d # e ", "http", nullptr, nullptr, "f", -2, "/ b ", " d ", " e"},
259
260 // Creative URLs missing key elements
261 {"", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
262 {" \t", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
263 {":foo.com/", "", nullptr, nullptr, "foo.com", -1, "/", nullptr, nullptr},
264 {":foo.com\\", "", nullptr, nullptr, "foo.com", -1, "\\", nullptr, nullptr},
265 {":", "", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
266 {":a", "", nullptr, nullptr, "a", -1, nullptr, nullptr, nullptr},
267 {":/", "", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
268 {":\\", "", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
269 {":#", "", nullptr, nullptr, nullptr, -1, nullptr, nullptr, ""},
270 {"#", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, ""},
271 {"#/", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, "/"},
272 {"#\\", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, "\\"},
273 {"#;?", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, ";?"},
274 {"?", nullptr,nullptr, nullptr, nullptr, -1, nullptr, "", nullptr},
275 {"/", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
276 {":23", "", nullptr, nullptr, "23", -1, nullptr, nullptr, nullptr},
277 {"/:23", "/", nullptr, nullptr, "23", -1, nullptr, nullptr, nullptr},
278 {"//", nullptr,nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
279 {"::", "", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
280 {"::23", "", nullptr, nullptr, nullptr, 23, nullptr, nullptr, nullptr},
281 {"foo://", "foo", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
282
283 // Username/passwords and things that look like them
284 {"http://a:b@c:29/d", "http", "a", "b", "c", 29, "/d", nullptr, nullptr},
285 {"http::@c:29", "http", "", "", "c", 29, nullptr, nullptr, nullptr},
286 // ... "]" in the password field isn't allowed, but we tolerate it here...
287 {"http://&a:foo(b]c@d:2/", "http", "&a", "foo(b]c", "d", 2, "/", nullptr, nullptr},
288 {"http://::@c@d:2", "http", "", ":@c", "d", 2, nullptr, nullptr, nullptr},
289 {"http://foo.com:b@d/", "http", "foo.com","b", "d", -1, "/", nullptr, nullptr},
290
291 {"http://foo.com/\\@", "http", nullptr, nullptr, "foo.com", -1, "/\\@", nullptr, nullptr},
292 {"http:\\\\foo.com\\", "http", nullptr, nullptr, "foo.com", -1, "\\", nullptr, nullptr},
293 {"http:\\\\a\\b:c\\[email protected]\\", "http", nullptr, nullptr, "a", -1, "\\b:c\\[email protected]\\", nullptr,nullptr},
294
295 // Tolerate different numbers of slashes.
296 {"foo:/", "foo", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
297 {"foo:/bar.com/", "foo", nullptr, nullptr, "bar.com", -1, "/", nullptr, nullptr},
298 {"foo://///////", "foo", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
299 {"foo://///////bar.com/", "foo", nullptr, nullptr, "bar.com", -1, "/", nullptr, nullptr},
300 {"foo:////://///", "foo", nullptr, nullptr, nullptr, -1, "/////", nullptr, nullptr},
301
302 // Raw file paths on Windows aren't handled by the parser.
303 {"c:/foo", "c", nullptr, nullptr, "foo", -1, nullptr, nullptr, nullptr},
304 {"//foo/bar", nullptr,nullptr, nullptr, "foo", -1, "/bar", nullptr, nullptr},
305
306 // Use the first question mark for the query and the ref.
307 {"http://foo/path;a??e#f#g", "http", nullptr, nullptr, "foo", -1, "/path;a", "?e", "f#g"},
308 {"http://foo/abcd?efgh?ijkl", "http", nullptr, nullptr, "foo", -1, "/abcd", "efgh?ijkl", nullptr},
309 {"http://foo/abcd#foo?bar", "http", nullptr, nullptr, "foo", -1, "/abcd", nullptr, "foo?bar"},
310
311 // IPv6, check also interesting uses of colons.
312 {"[61:24:74]:98", "[61", nullptr, nullptr, "24:74]", 98, nullptr, nullptr, nullptr},
313 {"http://[61:27]:98", "http", nullptr, nullptr, "[61:27]", 98, nullptr, nullptr, nullptr},
314 {"http:[61:27]/:foo", "http", nullptr, nullptr, "[61:27]", -1, "/:foo", nullptr, nullptr},
315 {"http://[1::2]:3:4", "http", nullptr, nullptr, "[1::2]:3", 4, nullptr, nullptr, nullptr},
316
317 // Partially-complete IPv6 literals, and related cases.
318 {"http://2001::1", "http", nullptr, nullptr, "2001:", 1, nullptr, nullptr, nullptr},
319 {"http://[2001::1", "http", nullptr, nullptr, "[2001::1", -1, nullptr, nullptr, nullptr},
320 {"http://2001::1]", "http", nullptr, nullptr, "2001::1]", -1, nullptr, nullptr, nullptr},
321 {"http://2001::1]:80", "http", nullptr, nullptr, "2001::1]", 80, nullptr, nullptr, nullptr},
322 {"http://[2001::1]", "http", nullptr, nullptr, "[2001::1]", -1, nullptr, nullptr, nullptr},
323 {"http://[2001::1]:80", "http", nullptr, nullptr, "[2001::1]", 80, nullptr, nullptr, nullptr},
324 {"http://[[::]]", "http", nullptr, nullptr, "[[::]]", -1, nullptr, nullptr, nullptr},
325
326 };
327 // clang-format on
328
TEST(URLParser,Standard)329 TEST(URLParser, Standard) {
330 // Declared outside for loop to try to catch cases in init() where we forget
331 // to reset something that is reset by the constructor.
332 Parsed parsed;
333 for (const auto& i : cases) {
334 const char* url = i.input;
335 ParseStandardURL(url, static_cast<int>(strlen(url)), &parsed);
336 URLParseCaseMatches(i, parsed);
337 }
338 }
339
340 // PathURL --------------------------------------------------------------------
341
342 // Various incarnations of path URLs.
343 // clang-format off
344 static PathURLParseCase path_cases[] = {
345 {"", nullptr, nullptr},
346 {":", "", nullptr},
347 {":/", "", "/"},
348 {"/", nullptr, "/"},
349 {" This is \\interesting// \t", nullptr, "This is \\interesting// \t"},
350 {"about:", "about", nullptr},
351 {"about:blank", "about", "blank"},
352 {" about: blank ", "about", " blank "},
353 {"javascript :alert(\"He:/l\\l#o?foo\"); ", "javascript ", "alert(\"He:/l\\l#o?foo\"); "},
354 };
355 // clang-format on
356
TEST(URLParser,PathURL)357 TEST(URLParser, PathURL) {
358 // Declared outside for loop to try to catch cases in init() where we forget
359 // to reset something that is reset by the constructor.
360 Parsed parsed;
361 for (size_t i = 0; i < std::size(path_cases); i++) {
362 const char* url = path_cases[i].input;
363 ParsePathURL(url, static_cast<int>(strlen(url)), false, &parsed);
364
365 EXPECT_TRUE(ComponentMatches(url, path_cases[i].scheme, parsed.scheme))
366 << i;
367 EXPECT_TRUE(ComponentMatches(url, path_cases[i].path, parsed.GetContent()))
368 << i;
369
370 // The remaining components are never used for path URLs.
371 ExpectInvalidComponent(parsed.username);
372 ExpectInvalidComponent(parsed.password);
373 ExpectInvalidComponent(parsed.host);
374 ExpectInvalidComponent(parsed.port);
375 }
376 }
377
378 // Various incarnations of file URLs.
379 // clang-format off
380 static URLParseCase file_cases[] = {
381 #ifdef WIN32
382 {"file:server", "file", nullptr, nullptr, "server", -1, nullptr, nullptr, nullptr},
383 {" file: server \t", "file", nullptr, nullptr, " server",-1, nullptr, nullptr, nullptr},
384 {"FiLe:c|", "FiLe", nullptr, nullptr, nullptr, -1, "c|", nullptr, nullptr},
385 {"FILE:/\\\\/server/file", "FILE", nullptr, nullptr, "server", -1, "/file", nullptr, nullptr},
386 {"file://server/", "file", nullptr, nullptr, "server", -1, "/", nullptr, nullptr},
387 {"file://localhost/c:/", "file", nullptr, nullptr, "localhost", -1, "/c:/", nullptr, nullptr},
388 {"file://127.0.0.1/c|\\", "file", nullptr, nullptr, "127.0.0.1", -1, "/c|\\", nullptr, nullptr},
389 {"file:/", "file", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
390 {"file:", "file", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
391 // If there is a Windows drive letter, treat any number of slashes as the
392 // path part.
393 {"file:c:\\fo\\b", "file", nullptr, nullptr, nullptr, -1, "c:\\fo\\b", nullptr, nullptr},
394 {"file:/c:\\foo/bar", "file", nullptr, nullptr, nullptr, -1, "/c:\\foo/bar",nullptr, nullptr},
395 {"file://c:/f\\b", "file", nullptr, nullptr, nullptr, -1, "/c:/f\\b", nullptr, nullptr},
396 {"file:///C:/foo", "file", nullptr, nullptr, nullptr, -1, "/C:/foo", nullptr, nullptr},
397 {"file://///\\/\\/c:\\f\\b", "file", nullptr, nullptr, nullptr, -1, "/c:\\f\\b", nullptr, nullptr},
398 // If there is not a drive letter, we should treat is as UNC EXCEPT for
399 // three slashes, which we treat as a Unix style path.
400 {"file:server/file", "file", nullptr, nullptr, "server", -1, "/file", nullptr, nullptr},
401 {"file:/server/file", "file", nullptr, nullptr, "server", -1, "/file", nullptr, nullptr},
402 {"file://server/file", "file", nullptr, nullptr, "server", -1, "/file", nullptr, nullptr},
403 {"file:///server/file", "file", nullptr, nullptr, nullptr, -1, "/server/file",nullptr, nullptr},
404 {"file://\\server/file", "file", nullptr, nullptr, nullptr, -1, "\\server/file",nullptr, nullptr},
405 {"file:////server/file", "file", nullptr, nullptr, "server", -1, "/file", nullptr, nullptr},
406 // Queries and refs are valid for file URLs as well.
407 {"file:///C:/foo.html?#", "file", nullptr, nullptr, nullptr, -1, "/C:/foo.html", "", ""},
408 {"file:///C:/foo.html?query=yes#ref", "file", nullptr, nullptr, nullptr, -1, "/C:/foo.html", "query=yes", "ref"},
409 #else // WIN32
410 // No slashes.
411 {"file:", "file", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
412 {"file:path", "file", nullptr, nullptr, nullptr, -1, "path", nullptr, nullptr},
413 {"file:path/", "file", nullptr, nullptr, nullptr, -1, "path/", nullptr, nullptr},
414 {"file:path/f.txt", "file", nullptr, nullptr, nullptr, -1, "path/f.txt", nullptr, nullptr},
415 // One slash.
416 {"file:/", "file", nullptr, nullptr, nullptr, -1, "/", nullptr, nullptr},
417 {"file:/path", "file", nullptr, nullptr, nullptr, -1, "/path", nullptr, nullptr},
418 {"file:/path/", "file", nullptr, nullptr, nullptr, -1, "/path/", nullptr, nullptr},
419 {"file:/path/f.txt", "file", nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
420 // Two slashes.
421 {"file://", "file", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
422 {"file://server", "file", nullptr, nullptr, "server", -1, nullptr, nullptr, nullptr},
423 {"file://server/", "file", nullptr, nullptr, "server", -1, "/", nullptr, nullptr},
424 {"file://server/f.txt", "file", nullptr, nullptr, "server", -1, "/f.txt", nullptr, nullptr},
425 // Three slashes.
426 {"file:///", "file", nullptr, nullptr, nullptr, -1, "/", nullptr, nullptr},
427 {"file:///path", "file", nullptr, nullptr, nullptr, -1, "/path", nullptr, nullptr},
428 {"file:///path/", "file", nullptr, nullptr, nullptr, -1, "/path/", nullptr, nullptr},
429 {"file:///path/f.txt", "file", nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
430 // More than three slashes.
431 {"file:////", "file", nullptr, nullptr, nullptr, -1, "/", nullptr, nullptr},
432 {"file:////path", "file", nullptr, nullptr, nullptr, -1, "/path", nullptr, nullptr},
433 {"file:////path/", "file", nullptr, nullptr, nullptr, -1, "/path/", nullptr, nullptr},
434 {"file:////path/f.txt", "file", nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
435 // Schemeless URLs
436 {"path/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "path/f.txt", nullptr, nullptr},
437 {"path:80/f.txt", "path", nullptr, nullptr, nullptr, -1, "80/f.txt", nullptr, nullptr},
438 {"path/f.txt:80", "path/f.txt",nullptr, nullptr, nullptr,-1,"80", nullptr, nullptr}, // Wrong.
439 {"/path/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
440 {"/path:80/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path:80/f.txt",nullptr, nullptr},
441 {"/path/f.txt:80", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt:80",nullptr, nullptr},
442 {"//server/f.txt", nullptr,nullptr, nullptr, "server", -1, "/f.txt", nullptr, nullptr},
443 {"//server:80/f.txt", nullptr,nullptr, nullptr, "server:80",-1, "/f.txt", nullptr, nullptr},
444 {"//server/f.txt:80", nullptr,nullptr, nullptr, "server", -1, "/f.txt:80", nullptr, nullptr},
445 {"///path/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
446 {"///path:80/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path:80/f.txt",nullptr, nullptr},
447 {"///path/f.txt:80", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt:80",nullptr, nullptr},
448 {"////path/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt", nullptr, nullptr},
449 {"////path:80/f.txt", nullptr,nullptr, nullptr, nullptr, -1, "/path:80/f.txt",nullptr, nullptr},
450 {"////path/f.txt:80", nullptr,nullptr, nullptr, nullptr, -1, "/path/f.txt:80",nullptr, nullptr},
451 // Queries and refs are valid for file URLs as well.
452 {"file:///foo.html?#", "file", nullptr, nullptr, nullptr, -1, "/foo.html", "", ""},
453 {"file:///foo.html?q=y#ref", "file", nullptr, nullptr, nullptr, -1, "/foo.html", "q=y", "ref"},
454 #endif // WIN32
455 };
456 // clang-format on
457
TEST(URLParser,ParseFileURL)458 TEST(URLParser, ParseFileURL) {
459 // Declared outside for loop to try to catch cases in init() where we forget
460 // to reset something that is reset by the construtor.
461 Parsed parsed;
462 for (const auto& file_case : file_cases) {
463 ParseFileURL(file_case.input, static_cast<int>(strlen(file_case.input)),
464 &parsed);
465 URLParseCaseMatches(file_case, parsed);
466 EXPECT_FALSE(parsed.has_opaque_path);
467 }
468 }
469
TEST(URLParser,ExtractFileName)470 TEST(URLParser, ExtractFileName) {
471 struct FileCase {
472 const char* input;
473 const char* expected;
474 } extract_cases[] = {
475 {"http://www.google.com", nullptr},
476 {"http://www.google.com/", ""},
477 {"http://www.google.com/search", "search"},
478 {"http://www.google.com/search/", ""},
479 {"http://www.google.com/foo/bar.html?baz=22", "bar.html"},
480 {"http://www.google.com/foo/bar.html#ref", "bar.html"},
481 {"http://www.google.com/search/;param", ""},
482 {"http://www.google.com/foo/bar.html;param#ref", "bar.html"},
483 {"http://www.google.com/foo/bar.html;foo;param#ref", "bar.html"},
484 {"http://www.google.com/foo/bar.html?query#ref", "bar.html"},
485 {"http://www.google.com/foo;/bar.html", "bar.html"},
486 {"http://www.google.com/foo;/", ""},
487 {"http://www.google.com/foo;", "foo"},
488 {"http://www.google.com/;", ""},
489 {"http://www.google.com/foo;bar;html", "foo"},
490 };
491
492 for (const auto& extract_case : extract_cases) {
493 const char* url = extract_case.input;
494 int len = static_cast<int>(strlen(url));
495
496 Parsed parsed;
497 ParseStandardURL(url, len, &parsed);
498
499 Component file_name;
500 ExtractFileName(url, parsed.path, &file_name);
501
502 EXPECT_TRUE(ComponentMatches(url, extract_case.expected, file_name));
503 }
504 }
505
506 // Returns true if the parameter with index |parameter| in the given URL's
507 // query string. The expected key can be NULL to indicate no such key index
508 // should exist. The parameter number is 1-based.
NthParameterIs(const char * url,int parameter,const char * expected_key,const char * expected_value)509 static bool NthParameterIs(const char* url,
510 int parameter,
511 const char* expected_key,
512 const char* expected_value) {
513 Parsed parsed;
514 ParseStandardURL(url, static_cast<int>(strlen(url)), &parsed);
515
516 Component query = parsed.query;
517
518 for (int i = 1; i <= parameter; i++) {
519 Component key, value;
520 if (!ExtractQueryKeyValue(url, &query, &key, &value)) {
521 if (parameter >= i && !expected_key)
522 return true; // Expected nonexistent key, got one.
523 return false; // Not enough keys.
524 }
525
526 if (i == parameter) {
527 if (!expected_key)
528 return false;
529
530 if (strncmp(&url[key.begin], expected_key, key.len) != 0)
531 return false;
532 if (strncmp(&url[value.begin], expected_value, value.len) != 0)
533 return false;
534 return true;
535 }
536 }
537 return expected_key == nullptr; // We didn't find that many parameters.
538 }
539
TEST(URLParser,ExtractQueryKeyValue)540 TEST(URLParser, ExtractQueryKeyValue) {
541 EXPECT_TRUE(NthParameterIs("http://www.google.com", 1, nullptr, nullptr));
542
543 // Basic case.
544 char a[] = "http://www.google.com?arg1=1&arg2=2&bar";
545 EXPECT_TRUE(NthParameterIs(a, 1, "arg1", "1"));
546 EXPECT_TRUE(NthParameterIs(a, 2, "arg2", "2"));
547 EXPECT_TRUE(NthParameterIs(a, 3, "bar", ""));
548 EXPECT_TRUE(NthParameterIs(a, 4, nullptr, nullptr));
549
550 // Empty param at the end.
551 char b[] = "http://www.google.com?foo=bar&";
552 EXPECT_TRUE(NthParameterIs(b, 1, "foo", "bar"));
553 EXPECT_TRUE(NthParameterIs(b, 2, nullptr, nullptr));
554
555 // Empty param at the beginning.
556 char c[] = "http://www.google.com?&foo=bar";
557 EXPECT_TRUE(NthParameterIs(c, 1, "", ""));
558 EXPECT_TRUE(NthParameterIs(c, 2, "foo", "bar"));
559 EXPECT_TRUE(NthParameterIs(c, 3, nullptr, nullptr));
560
561 // Empty key with value.
562 char d[] = "http://www.google.com?=foo";
563 EXPECT_TRUE(NthParameterIs(d, 1, "", "foo"));
564 EXPECT_TRUE(NthParameterIs(d, 2, nullptr, nullptr));
565
566 // Empty value with key.
567 char e[] = "http://www.google.com?foo=";
568 EXPECT_TRUE(NthParameterIs(e, 1, "foo", ""));
569 EXPECT_TRUE(NthParameterIs(e, 2, nullptr, nullptr));
570
571 // Empty key and values.
572 char f[] = "http://www.google.com?&&==&=";
573 EXPECT_TRUE(NthParameterIs(f, 1, "", ""));
574 EXPECT_TRUE(NthParameterIs(f, 2, "", ""));
575 EXPECT_TRUE(NthParameterIs(f, 3, "", "="));
576 EXPECT_TRUE(NthParameterIs(f, 4, "", ""));
577 EXPECT_TRUE(NthParameterIs(f, 5, nullptr, nullptr));
578 }
579
580 // MailtoURL --------------------------------------------------------------------
581
582 // clang-format off
583 static MailtoURLParseCase mailto_cases[] = {
584 //|input |scheme |path |query
585 {"mailto:[email protected]", "mailto", "[email protected]", nullptr},
586 {" mailto: to \t", "mailto", " to", nullptr},
587 {"mailto:addr1%2C%20addr2 ", "mailto", "addr1%2C%20addr2", nullptr},
588 {"Mailto:addr1, addr2 ", "Mailto", "addr1, addr2", nullptr},
589 {"mailto:addr1:addr2 ", "mailto", "addr1:addr2", nullptr},
590 {"mailto:?to=addr1,addr2", "mailto", nullptr, "to=addr1,addr2"},
591 {"mailto:?to=addr1%2C%20addr2", "mailto", nullptr, "to=addr1%2C%20addr2"},
592 {"mailto:addr1?to=addr2", "mailto", "addr1", "to=addr2"},
593 {"mailto:?body=#foobar#", "mailto", nullptr, "body=#foobar#",},
594 {"mailto:#?body=#foobar#", "mailto", "#", "body=#foobar#"},
595 };
596 // clang-format on
597
TEST(URLParser,MailtoUrl)598 TEST(URLParser, MailtoUrl) {
599 // Declared outside for loop to try to catch cases in init() where we forget
600 // to reset something that is reset by the constructor.
601 for (const auto& mailto_case : mailto_cases) {
602 const char* url = mailto_case.input;
603 Parsed parsed = ParseMailtoURL(url);
604 int port = ParsePort(url, parsed.port);
605
606 EXPECT_TRUE(ComponentMatches(url, mailto_case.scheme, parsed.scheme));
607 EXPECT_TRUE(ComponentMatches(url, mailto_case.path, parsed.path));
608 EXPECT_TRUE(ComponentMatches(url, mailto_case.query, parsed.query));
609 EXPECT_EQ(PORT_UNSPECIFIED, port);
610 EXPECT_FALSE(parsed.has_opaque_path);
611
612 // The remaining components are never used for mailto URLs.
613 ExpectInvalidComponent(parsed.username);
614 ExpectInvalidComponent(parsed.password);
615 ExpectInvalidComponent(parsed.port);
616 ExpectInvalidComponent(parsed.ref);
617 }
618 }
619
620 // Various incarnations of filesystem URLs.
621 static FileSystemURLParseCase filesystem_cases[] = {
622 // Regular URL with all the parts
623 {"filesystem:http://user:pass@foo:21/temporary/bar;par?b#c", "http", "user",
624 "pass", "foo", 21, "/temporary", "/bar;par", "b", "c"},
625 {"filesystem:https://foo/persistent/bar;par/", "https", nullptr, nullptr,
626 "foo", -1, "/persistent", "/bar;par/", nullptr, nullptr},
627 {"filesystem:file:///persistent/bar;par/", "file", nullptr, nullptr,
628 nullptr, -1, "/persistent", "/bar;par/", nullptr, nullptr},
629 {"filesystem:file:///persistent/bar;par/?query#ref", "file", nullptr,
630 nullptr, nullptr, -1, "/persistent", "/bar;par/", "query", "ref"},
631 {"filesystem:file:///persistent", "file", nullptr, nullptr, nullptr, -1,
632 "/persistent", "", nullptr, nullptr},
633 };
634
TEST(URLParser,FileSystemURL)635 TEST(URLParser, FileSystemURL) {
636 // Declared outside for loop to try to catch cases in init() where we forget
637 // to reset something that is reset by the constructor.
638 for (const auto& filesystem_case : filesystem_cases) {
639 const char* url = filesystem_case.input;
640 Parsed parsed = ParseFileSystemURL(url);
641
642 EXPECT_TRUE(ComponentMatches(url, "filesystem", parsed.scheme));
643 EXPECT_EQ(!filesystem_case.inner_scheme, !parsed.inner_parsed());
644 // Only check the inner_parsed if there is one.
645 if (parsed.inner_parsed()) {
646 EXPECT_TRUE(ComponentMatches(url, filesystem_case.inner_scheme,
647 parsed.inner_parsed()->scheme));
648 EXPECT_TRUE(ComponentMatches(url, filesystem_case.inner_username,
649 parsed.inner_parsed()->username));
650 EXPECT_TRUE(ComponentMatches(url, filesystem_case.inner_password,
651 parsed.inner_parsed()->password));
652 EXPECT_TRUE(ComponentMatches(url, filesystem_case.inner_host,
653 parsed.inner_parsed()->host));
654 int port = ParsePort(url, parsed.inner_parsed()->port);
655 EXPECT_EQ(filesystem_case.inner_port, port);
656
657 // The remaining components are never used for filesystem URLs.
658 ExpectInvalidComponent(parsed.inner_parsed()->query);
659 ExpectInvalidComponent(parsed.inner_parsed()->ref);
660 }
661
662 EXPECT_TRUE(ComponentMatches(url, filesystem_case.path, parsed.path));
663 EXPECT_TRUE(ComponentMatches(url, filesystem_case.query, parsed.query));
664 EXPECT_TRUE(ComponentMatches(url, filesystem_case.ref, parsed.ref));
665 EXPECT_FALSE(parsed.has_opaque_path);
666
667 // The remaining components are never used for filesystem URLs.
668 ExpectInvalidComponent(parsed.username);
669 ExpectInvalidComponent(parsed.password);
670 ExpectInvalidComponent(parsed.host);
671 ExpectInvalidComponent(parsed.port);
672 }
673 }
674
675 // Non-special URLs which don't have an opaque path.
676 static URLParseCase non_special_cases[] = {
677 {"git://user:pass@foo:21/bar;par?b#c", "git", "user", "pass", "foo", 21,
678 "/bar;par", "b", "c"},
679 {"git://host", "git", nullptr, nullptr, "host", -1, nullptr, nullptr,
680 nullptr},
681 {"git://host/a/../b", "git", nullptr, nullptr, "host", -1, "/a/../b",
682 nullptr, nullptr},
683 {"git://host/a b", "git", nullptr, nullptr, "host", -1, "/a b", nullptr,
684 nullptr},
685 {"git://ho\\st/", "git", nullptr, nullptr, "ho\\st", -1, "/", nullptr,
686 nullptr},
687 // Empty users
688 {"git://@host", "git", "", nullptr, "host", -1, nullptr, nullptr, nullptr},
689 // Empty user and invalid host. "git://@" is an invalid URL.
690 {"git://@", "git", "", nullptr, nullptr, -1, nullptr, nullptr, nullptr},
691 // Invalid host and non-empty port. "git://:80" is an invalid URL.
692 {"git://:80", "git", nullptr, nullptr, nullptr, 80, nullptr, nullptr,
693 nullptr},
694 // Empty host cases
695 {"git://", "git", nullptr, nullptr, "", -1, nullptr, nullptr, nullptr},
696 {"git:///", "git", nullptr, nullptr, "", -1, "/", nullptr, nullptr},
697 {"git:////", "git", nullptr, nullptr, "", -1, "//", nullptr, nullptr},
698 // Null host cases
699 {"git:/", "git", nullptr, nullptr, nullptr, -1, "/", nullptr, nullptr},
700 {"git:/trailing-space ", "git", nullptr, nullptr, nullptr, -1,
701 "/trailing-space", nullptr, nullptr},
702 };
703
TEST(URLParser,NonSpecial)704 TEST(URLParser, NonSpecial) {
705 // Declared outside for loop to try to catch cases in init() where we forget
706 // to reset something that is reset by the constructor.
707 Parsed parsed;
708 for (const auto& i : non_special_cases) {
709 const char* url = i.input;
710 ParseNonSpecialURL(url, static_cast<int>(strlen(url)), &parsed);
711 URLParseCaseMatches(i, parsed);
712 EXPECT_FALSE(parsed.has_opaque_path) << "url: " << url;
713 }
714 }
715
716 // Non-special URLs which have an opaque path.
717 static URLParseCase non_special_opaque_path_cases[] = {
718 {"git:", "git", nullptr, nullptr, nullptr, -1, nullptr, nullptr, nullptr},
719 {"git:opaque", "git", nullptr, nullptr, nullptr, -1, "opaque", nullptr,
720 nullptr},
721 {"git:opaque?a=b#c", "git", nullptr, nullptr, nullptr, -1, "opaque", "a=b",
722 "c"},
723 {"git: o p a q u e ", "git", nullptr, nullptr, nullptr, -1, " o p a q u e",
724 nullptr, nullptr},
725 {"git:opa\\que", "git", nullptr, nullptr, nullptr, -1, "opa\\que", nullptr,
726 nullptr},
727 };
728
TEST(URLParser,NonSpecialOpaquePath)729 TEST(URLParser, NonSpecialOpaquePath) {
730 // Declared outside for loop to try to catch cases in init() where we forget
731 // to reset something that is reset by the constructor.
732 Parsed parsed;
733 for (const auto& i : non_special_opaque_path_cases) {
734 const char* url = i.input;
735 ParseNonSpecialURL(url, static_cast<int>(strlen(url)), &parsed);
736 URLParseCaseMatches(i, parsed);
737 EXPECT_TRUE(parsed.has_opaque_path) << "url: " << url;
738 }
739 }
740
741 } // namespace
742 } // namespace url
743