1 // Copyright (c) 2011 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
7 #include "base/basictypes.h"
8 #include "base/file_util.h"
9 #include "base/path_service.h"
10 #include "base/strings/string_util.h"
11 #include "base/strings/utf_string_conversions.h"
12 #include "chrome/common/chrome_paths.h"
13 #include "chrome/common/net/url_fixer_upper.h"
14 #include "net/base/net_util.h"
15 #include "testing/gtest/include/gtest/gtest.h"
17 #include "url/url_parse.h"
20 class URLFixerUpperTest
: public testing::Test
{
26 std::ostream
& operator<<(std::ostream
& os
, const Component
& part
) {
27 return os
<< "(begin=" << part
.begin
<< ", len=" << part
.len
<< ")";
30 } // namespace url_parse
33 const std::string input
;
34 const std::string result
;
35 const url_parse::Component scheme
;
36 const url_parse::Component username
;
37 const url_parse::Component password
;
38 const url_parse::Component host
;
39 const url_parse::Component port
;
40 const url_parse::Component path
;
41 const url_parse::Component query
;
42 const url_parse::Component ref
;
45 static const SegmentCase segment_cases
[] = {
46 { "http://www.google.com/", "http",
47 url_parse::Component(0, 4), // scheme
48 url_parse::Component(), // username
49 url_parse::Component(), // password
50 url_parse::Component(7, 14), // host
51 url_parse::Component(), // port
52 url_parse::Component(21, 1), // path
53 url_parse::Component(), // query
54 url_parse::Component(), // ref
56 { "aBoUt:vErSiOn", "about",
57 url_parse::Component(0, 5), // scheme
58 url_parse::Component(), // username
59 url_parse::Component(), // password
60 url_parse::Component(6, 7), // host
61 url_parse::Component(), // port
62 url_parse::Component(), // path
63 url_parse::Component(), // query
64 url_parse::Component(), // ref
66 { "about:host/path?query#ref", "about",
67 url_parse::Component(0, 5), // scheme
68 url_parse::Component(), // username
69 url_parse::Component(), // password
70 url_parse::Component(6, 4), // host
71 url_parse::Component(), // port
72 url_parse::Component(10, 5), // path
73 url_parse::Component(16, 5), // query
74 url_parse::Component(22, 3), // ref
76 { "about://host/path?query#ref", "about",
77 url_parse::Component(0, 5), // scheme
78 url_parse::Component(), // username
79 url_parse::Component(), // password
80 url_parse::Component(8, 4), // host
81 url_parse::Component(), // port
82 url_parse::Component(12, 5), // path
83 url_parse::Component(18, 5), // query
84 url_parse::Component(24, 3), // ref
86 { "chrome:host/path?query#ref", "chrome",
87 url_parse::Component(0, 6), // scheme
88 url_parse::Component(), // username
89 url_parse::Component(), // password
90 url_parse::Component(7, 4), // host
91 url_parse::Component(), // port
92 url_parse::Component(11, 5), // path
93 url_parse::Component(17, 5), // query
94 url_parse::Component(23, 3), // ref
96 { "chrome://host/path?query#ref", "chrome",
97 url_parse::Component(0, 6), // scheme
98 url_parse::Component(), // username
99 url_parse::Component(), // password
100 url_parse::Component(9, 4), // host
101 url_parse::Component(), // port
102 url_parse::Component(13, 5), // path
103 url_parse::Component(19, 5), // query
104 url_parse::Component(25, 3), // ref
106 { " www.google.com:124?foo#", "http",
107 url_parse::Component(), // scheme
108 url_parse::Component(), // username
109 url_parse::Component(), // password
110 url_parse::Component(4, 14), // host
111 url_parse::Component(19, 3), // port
112 url_parse::Component(), // path
113 url_parse::Component(23, 3), // query
114 url_parse::Component(27, 0), // ref
116 { "user@www.google.com", "http",
117 url_parse::Component(), // scheme
118 url_parse::Component(0, 4), // username
119 url_parse::Component(), // password
120 url_parse::Component(5, 14), // host
121 url_parse::Component(), // port
122 url_parse::Component(), // path
123 url_parse::Component(), // query
124 url_parse::Component(), // ref
126 { "ftp:/user:P:a$$Wd@..ftp.google.com...::23///pub?foo#bar", "ftp",
127 url_parse::Component(0, 3), // scheme
128 url_parse::Component(5, 4), // username
129 url_parse::Component(10, 7), // password
130 url_parse::Component(18, 20), // host
131 url_parse::Component(39, 2), // port
132 url_parse::Component(41, 6), // path
133 url_parse::Component(48, 3), // query
134 url_parse::Component(52, 3), // ref
136 { "[2001:db8::1]/path", "http",
137 url_parse::Component(), // scheme
138 url_parse::Component(), // username
139 url_parse::Component(), // password
140 url_parse::Component(0, 13), // host
141 url_parse::Component(), // port
142 url_parse::Component(13, 5), // path
143 url_parse::Component(), // query
144 url_parse::Component(), // ref
147 url_parse::Component(), // scheme
148 url_parse::Component(), // username
149 url_parse::Component(), // password
150 url_parse::Component(0, 5), // host
151 url_parse::Component(), // port
152 url_parse::Component(), // path
153 url_parse::Component(), // query
154 url_parse::Component(), // ref
156 // Incomplete IPv6 addresses (will not canonicalize).
157 { "[2001:4860:", "http",
158 url_parse::Component(), // scheme
159 url_parse::Component(), // username
160 url_parse::Component(), // password
161 url_parse::Component(0, 11), // host
162 url_parse::Component(), // port
163 url_parse::Component(), // path
164 url_parse::Component(), // query
165 url_parse::Component(), // ref
167 { "[2001:4860:/foo", "http",
168 url_parse::Component(), // scheme
169 url_parse::Component(), // username
170 url_parse::Component(), // password
171 url_parse::Component(0, 11), // host
172 url_parse::Component(), // port
173 url_parse::Component(11, 4), // path
174 url_parse::Component(), // query
175 url_parse::Component(), // ref
177 { "http://:b005::68]", "http",
178 url_parse::Component(0, 4), // scheme
179 url_parse::Component(), // username
180 url_parse::Component(), // password
181 url_parse::Component(7, 10), // host
182 url_parse::Component(), // port
183 url_parse::Component(), // path
184 url_parse::Component(), // query
185 url_parse::Component(), // ref
187 // Can't do anything useful with this.
189 url_parse::Component(0, 0), // scheme
190 url_parse::Component(), // username
191 url_parse::Component(), // password
192 url_parse::Component(), // host
193 url_parse::Component(), // port
194 url_parse::Component(), // path
195 url_parse::Component(), // query
196 url_parse::Component(), // ref
200 TEST(URLFixerUpperTest
, SegmentURL
) {
202 url_parse::Parsed parts
;
204 for (size_t i
= 0; i
< arraysize(segment_cases
); ++i
) {
205 SegmentCase value
= segment_cases
[i
];
206 result
= URLFixerUpper::SegmentURL(value
.input
, &parts
);
207 EXPECT_EQ(value
.result
, result
);
208 EXPECT_EQ(value
.scheme
, parts
.scheme
);
209 EXPECT_EQ(value
.username
, parts
.username
);
210 EXPECT_EQ(value
.password
, parts
.password
);
211 EXPECT_EQ(value
.host
, parts
.host
);
212 EXPECT_EQ(value
.port
, parts
.port
);
213 EXPECT_EQ(value
.path
, parts
.path
);
214 EXPECT_EQ(value
.query
, parts
.query
);
215 EXPECT_EQ(value
.ref
, parts
.ref
);
219 // Creates a file and returns its full name as well as the decomposed
221 // full_path = "c:\foo\bar.txt"
223 // file_name = "bar.txt"
224 static bool MakeTempFile(const base::FilePath
& dir
,
225 const base::FilePath
& file_name
,
226 base::FilePath
* full_path
) {
227 *full_path
= dir
.Append(file_name
);
228 return file_util::WriteFile(*full_path
, "", 0) == 0;
231 // Returns true if the given URL is a file: URL that matches the given file
232 static bool IsMatchingFileURL(const std::string
& url
,
233 const base::FilePath
& full_file_path
) {
234 if (url
.length() <= 8)
236 if (std::string("file:///") != url
.substr(0, 8))
237 return false; // no file:/// prefix
238 if (url
.find('\\') != std::string::npos
)
239 return false; // contains backslashes
241 base::FilePath derived_path
;
242 net::FileURLToFilePath(GURL(url
), &derived_path
);
244 return base::FilePath::CompareEqualIgnoreCase(derived_path
.value(),
245 full_file_path
.value());
249 const std::string input
;
250 const std::string desired_tld
;
251 const std::string output
;
253 {"www.google.com", "", "http://www.google.com/"},
254 {" www.google.com ", "", "http://www.google.com/"},
255 {" foo.com/asdf bar", "", "http://foo.com/asdf%20%20bar"},
256 {"..www.google.com..", "", "http://www.google.com./"},
257 {"http://......", "", "http://....../"},
258 {"http://host.com:ninety-two/", "", "http://host.com:ninety-two/"},
259 {"http://host.com:ninety-two?foo", "", "http://host.com:ninety-two/?foo"},
260 {"google.com:123", "", "http://google.com:123/"},
261 {"about:", "", "chrome://version/"},
262 {"about:foo", "", "chrome://foo/"},
263 {"about:version", "", "chrome://version/"},
264 {"about:usr:pwd@hst/pth?qry#ref", "", "chrome://usr:pwd@hst/pth?qry#ref"},
265 {"about://usr:pwd@hst/pth?qry#ref", "", "chrome://usr:pwd@hst/pth?qry#ref"},
266 {"chrome:usr:pwd@hst/pth?qry#ref", "", "chrome://usr:pwd@hst/pth?qry#ref"},
267 {"chrome://usr:pwd@hst/pth?qry#ref", "", "chrome://usr:pwd@hst/pth?qry#ref"},
268 {"www:123", "", "http://www:123/"},
269 {" www:123", "", "http://www:123/"},
270 {"www.google.com?foo", "", "http://www.google.com/?foo"},
271 {"www.google.com#foo", "", "http://www.google.com/#foo"},
272 {"www.google.com?", "", "http://www.google.com/?"},
273 {"www.google.com#", "", "http://www.google.com/#"},
274 {"www.google.com:123?foo#bar", "", "http://www.google.com:123/?foo#bar"},
275 {"user@www.google.com", "", "http://user@www.google.com/"},
276 {"\xE6\xB0\xB4.com" , "", "http://xn--1rw.com/"},
277 // It would be better if this next case got treated as http, but I don't see
278 // a clean way to guess this isn't the new-and-exciting "user" scheme.
279 {"user:passwd@www.google.com:8080/", "", "user:passwd@www.google.com:8080/"},
280 // {"file:///c:/foo/bar%20baz.txt", "", "file:///C:/foo/bar%20baz.txt"},
281 {"ftp.google.com", "", "ftp://ftp.google.com/"},
282 {" ftp.google.com", "", "ftp://ftp.google.com/"},
283 {"FTP.GooGle.com", "", "ftp://ftp.google.com/"},
284 {"ftpblah.google.com", "", "http://ftpblah.google.com/"},
285 {"ftp", "", "http://ftp/"},
286 {"google.ftp.com", "", "http://google.ftp.com/"},
287 // URLs which end with 0x85 (NEL in ISO-8859).
288 { "http://google.com/search?q=\xd0\x85", "",
289 "http://google.com/search?q=%D0%85"
291 { "http://google.com/search?q=\xec\x97\x85", "",
292 "http://google.com/search?q=%EC%97%85"
294 { "http://google.com/search?q=\xf0\x90\x80\x85", "",
295 "http://google.com/search?q=%F0%90%80%85"
297 // URLs which end with 0xA0 (non-break space in ISO-8859).
298 { "http://google.com/search?q=\xd0\xa0", "",
299 "http://google.com/search?q=%D0%A0"
301 { "http://google.com/search?q=\xec\x97\xa0", "",
302 "http://google.com/search?q=%EC%97%A0"
304 { "http://google.com/search?q=\xf0\x90\x80\xa0", "",
305 "http://google.com/search?q=%F0%90%80%A0"
307 // URLs containing IPv6 literals.
308 {"[2001:db8::2]", "", "http://[2001:db8::2]/"},
309 {"[::]:80", "", "http://[::]/"},
310 {"[::]:80/path", "", "http://[::]/path"},
311 {"[::]:180/path", "", "http://[::]:180/path"},
312 // TODO(pmarks): Maybe we should parse bare IPv6 literals someday.
314 // Semicolon as scheme separator for standard schemes.
315 {"http;//www.google.com/", "", "http://www.google.com/"},
316 {"about;chrome", "", "chrome://chrome/"},
317 // Semicolon left as-is for non-standard schemes.
318 {"whatsup;//fool", "", "whatsup://fool"},
319 // Semicolon left as-is in URL itself.
320 {"http://host/port?query;moar", "", "http://host/port?query;moar"},
321 // Fewer slashes than expected.
322 {"http;www.google.com/", "", "http://www.google.com/"},
323 {"http;/www.google.com/", "", "http://www.google.com/"},
324 // Semicolon at start.
325 {";http://www.google.com/", "", "http://%3Bhttp//www.google.com/"},
328 TEST(URLFixerUpperTest
, FixupURL
) {
329 for (size_t i
= 0; i
< arraysize(fixup_cases
); ++i
) {
330 FixupCase value
= fixup_cases
[i
];
331 EXPECT_EQ(value
.output
, URLFixerUpper::FixupURL(value
.input
,
332 value
.desired_tld
).possibly_invalid_spec())
333 << "input: " << value
.input
;
336 // Check the TLD-appending functionality
337 FixupCase tld_cases
[] = {
338 {"google", "com", "http://www.google.com/"},
339 {"google.", "com", "http://www.google.com/"},
340 {"google..", "com", "http://www.google.com/"},
341 {".google", "com", "http://www.google.com/"},
342 {"www.google", "com", "http://www.google.com/"},
343 {"google.com", "com", "http://google.com/"},
344 {"http://google", "com", "http://www.google.com/"},
345 {"..google..", "com", "http://www.google.com/"},
346 {"http://www.google", "com", "http://www.google.com/"},
347 {"9999999999999999", "com", "http://www.9999999999999999.com/"},
348 {"google/foo", "com", "http://www.google.com/foo"},
349 {"google.com/foo", "com", "http://google.com/foo"},
350 {"google/?foo=.com", "com", "http://www.google.com/?foo=.com"},
351 {"www.google/?foo=www.", "com", "http://www.google.com/?foo=www."},
352 {"google.com/?foo=.com", "com", "http://google.com/?foo=.com"},
353 {"http://www.google.com", "com", "http://www.google.com/"},
354 {"google:123", "com", "http://www.google.com:123/"},
355 {"http://google:123", "com", "http://www.google.com:123/"},
357 for (size_t i
= 0; i
< arraysize(tld_cases
); ++i
) {
358 FixupCase value
= tld_cases
[i
];
359 EXPECT_EQ(value
.output
, URLFixerUpper::FixupURL(value
.input
,
360 value
.desired_tld
).possibly_invalid_spec());
364 // Test different types of file inputs to URIFixerUpper::FixupURL. This
365 // doesn't go into the nice array of fixups above since the file input
367 TEST(URLFixerUpperTest
, FixupFile
) {
368 // this "original" filename is the one we tweak to get all the variations
370 base::FilePath original
;
371 ASSERT_TRUE(PathService::Get(chrome::DIR_APP
, &dir
));
372 ASSERT_TRUE(MakeTempFile(
374 base::FilePath(FILE_PATH_LITERAL("url fixer upper existing file.txt")),
378 GURL
golden(net::FilePathToFileURL(original
));
380 // c:\foo\bar.txt -> file:///c:/foo/bar.txt (basic)
382 GURL
fixedup(URLFixerUpper::FixupURL(base::WideToUTF8(original
.value()),
384 #elif defined(OS_POSIX)
385 GURL
fixedup(URLFixerUpper::FixupURL(original
.value(), std::string()));
387 EXPECT_EQ(golden
, fixedup
);
389 // TODO(port): Make some equivalent tests for posix.
391 // c|/foo\bar.txt -> file:///c:/foo/bar.txt (pipe allowed instead of colon)
392 std::string
cur(base::WideToUTF8(original
.value()));
393 EXPECT_EQ(':', cur
[1]);
395 EXPECT_EQ(golden
, URLFixerUpper::FixupURL(cur
, std::string()));
397 FixupCase file_cases
[] = {
398 {"c:\\This%20is a non-existent file.txt", "",
399 "file:///C:/This%2520is%20a%20non-existent%20file.txt"},
401 // \\foo\bar.txt -> file://foo/bar.txt
402 // UNC paths, this file won't exist, but since there are no escapes, it
403 // should be returned just converted to a file: URL.
404 {"\\\\SomeNonexistentHost\\foo\\bar.txt", "",
405 "file://somenonexistenthost/foo/bar.txt"},
406 // We do this strictly, like IE8, which only accepts this form using
407 // backslashes and not forward ones. Turning "//foo" into "http" matches
408 // Firefox and IE, silly though it may seem (it falls out of adding "http"
409 // as the default protocol if you haven't entered one).
410 {"//SomeNonexistentHost\\foo/bar.txt", "",
411 "http://somenonexistenthost/foo/bar.txt"},
412 {"file:///C:/foo/bar", "", "file:///C:/foo/bar"},
414 // Much of the work here comes from GURL's canonicalization stage.
415 {"file://C:/foo/bar", "", "file:///C:/foo/bar"},
416 {"file:c:", "", "file:///C:/"},
417 {"file:c:WINDOWS", "", "file:///C:/WINDOWS"},
418 {"file:c|Program Files", "", "file:///C:/Program%20Files"},
419 {"file:/file", "", "file://file/"},
420 {"file:////////c:\\foo", "", "file:///C:/foo"},
421 {"file://server/folder/file", "", "file://server/folder/file"},
423 // These are fixups we don't do, but could consider:
425 // {"file:///foo:/bar", "", "file://foo/bar"},
426 // {"file:/\\/server\\folder/file", "", "file://server/folder/file"},
428 #elif defined(OS_POSIX)
430 #if defined(OS_MACOSX)
431 #define HOME "/Users/"
433 #define HOME "/home/"
435 URLFixerUpper::home_directory_override
= "/foo";
436 FixupCase file_cases
[] = {
437 // File URLs go through GURL, which tries to escape intelligently.
438 {"/This%20is a non-existent file.txt", "",
439 "file:///This%2520is%20a%20non-existent%20file.txt"},
440 // A plain "/" refers to the root.
444 // These rely on the above home_directory_override.
450 // References to other users' homedirs.
452 "file://" HOME
"foo"},
454 "file://" HOME
"x/blah"},
457 for (size_t i
= 0; i
< arraysize(file_cases
); i
++) {
458 EXPECT_EQ(file_cases
[i
].output
, URLFixerUpper::FixupURL(file_cases
[i
].input
,
459 file_cases
[i
].desired_tld
).possibly_invalid_spec());
462 EXPECT_TRUE(base::DeleteFile(original
, false));
465 TEST(URLFixerUpperTest
, FixupRelativeFile
) {
466 base::FilePath full_path
, dir
;
467 base::FilePath
file_part(
468 FILE_PATH_LITERAL("url_fixer_upper_existing_file.txt"));
469 ASSERT_TRUE(PathService::Get(chrome::DIR_APP
, &dir
));
470 ASSERT_TRUE(MakeTempFile(dir
, file_part
, &full_path
));
471 full_path
= base::MakeAbsoluteFilePath(full_path
);
472 ASSERT_FALSE(full_path
.empty());
474 // make sure we pass through good URLs
475 for (size_t i
= 0; i
< arraysize(fixup_cases
); ++i
) {
476 FixupCase value
= fixup_cases
[i
];
477 base::FilePath input
= base::FilePath::FromUTF8Unsafe(value
.input
);
478 EXPECT_EQ(value
.output
,
479 URLFixerUpper::FixupRelativeFile(dir
, input
).possibly_invalid_spec());
482 // make sure the existing file got fixed-up to a file URL, and that there
483 // are no backslashes
484 EXPECT_TRUE(IsMatchingFileURL(URLFixerUpper::FixupRelativeFile(dir
,
485 file_part
).possibly_invalid_spec(), full_path
));
486 EXPECT_TRUE(base::DeleteFile(full_path
, false));
488 // create a filename we know doesn't exist and make sure it doesn't get
489 // fixed up to a file URL
490 base::FilePath
nonexistent_file(
491 FILE_PATH_LITERAL("url_fixer_upper_nonexistent_file.txt"));
492 std::string
fixedup(URLFixerUpper::FixupRelativeFile(dir
,
493 nonexistent_file
).possibly_invalid_spec());
494 EXPECT_NE(std::string("file:///"), fixedup
.substr(0, 8));
495 EXPECT_FALSE(IsMatchingFileURL(fixedup
, nonexistent_file
));
497 // make a subdir to make sure relative paths with directories work, also
499 // "app_dir\url fixer-upper dir\url fixer-upper existing file.txt"
500 base::FilePath
sub_dir(FILE_PATH_LITERAL("url fixer-upper dir"));
501 base::FilePath
sub_file(
502 FILE_PATH_LITERAL("url fixer-upper existing file.txt"));
503 base::FilePath new_dir
= dir
.Append(sub_dir
);
504 base::CreateDirectory(new_dir
);
505 ASSERT_TRUE(MakeTempFile(new_dir
, sub_file
, &full_path
));
506 full_path
= base::MakeAbsoluteFilePath(full_path
);
507 ASSERT_FALSE(full_path
.empty());
509 // test file in the subdir
510 base::FilePath relative_file
= sub_dir
.Append(sub_file
);
511 EXPECT_TRUE(IsMatchingFileURL(URLFixerUpper::FixupRelativeFile(dir
,
512 relative_file
).possibly_invalid_spec(), full_path
));
514 // test file in the subdir with different slashes and escaping.
515 base::FilePath::StringType relative_file_str
= sub_dir
.value() +
516 FILE_PATH_LITERAL("/") + sub_file
.value();
517 ReplaceSubstringsAfterOffset(&relative_file_str
, 0,
518 FILE_PATH_LITERAL(" "), FILE_PATH_LITERAL("%20"));
519 EXPECT_TRUE(IsMatchingFileURL(URLFixerUpper::FixupRelativeFile(dir
,
520 base::FilePath(relative_file_str
)).possibly_invalid_spec(), full_path
));
522 // test relative directories and duplicate slashes
523 // (should resolve to the same file as above)
524 relative_file_str
= sub_dir
.value() + FILE_PATH_LITERAL("/../") +
525 sub_dir
.value() + FILE_PATH_LITERAL("///./") + sub_file
.value();
526 EXPECT_TRUE(IsMatchingFileURL(URLFixerUpper::FixupRelativeFile(dir
,
527 base::FilePath(relative_file_str
)).possibly_invalid_spec(), full_path
));
529 // done with the subdir
530 EXPECT_TRUE(base::DeleteFile(full_path
, false));
531 EXPECT_TRUE(base::DeleteFile(new_dir
, true));
533 // Test that an obvious HTTP URL isn't accidentally treated as an absolute
534 // file path (on account of system-specific craziness).
535 base::FilePath empty_path
;
536 base::FilePath
http_url_path(FILE_PATH_LITERAL("http://../"));
537 EXPECT_TRUE(URLFixerUpper::FixupRelativeFile(
538 empty_path
, http_url_path
).SchemeIs("http"));