url/gurl_unittest.cc

   1 // Copyright 2013 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.
   4
   5 #include "testing/gtest/include/gtest/gtest.h"
   6 #include "url/gurl.h"
   7 #include "url/url_canon.h"
   8 #include "url/url_test_utils.h"
   9
  10 // Some implementations of base/basictypes.h may define ARRAYSIZE.
  11 // If it's not defined, we define it to the ARRAYSIZE_UNSAFE macro
  12 // which is in our version of basictypes.h.
  13 #ifndef ARRAYSIZE
  14 #define ARRAYSIZE ARRAYSIZE_UNSAFE
  15 #endif
  16
  17 using url_test_utils::WStringToUTF16;
  18 using url_test_utils::ConvertUTF8ToUTF16;
  19
  20 namespace {
  21
  22 template<typename CHAR>
  23 void SetupReplacement(void (url_canon::Replacements<CHAR>::*func)(const CHAR*,
  24                           const url_parse::Component&),
  25                       url_canon::Replacements<CHAR>* replacements,
  26                       const CHAR* str) {
  27   if (str) {
  28     url_parse::Component comp;
  29     if (str[0])
  30       comp.len = static_cast<int>(strlen(str));
  31     (replacements->*func)(str, comp);
  32   }
  33 }
  34
  35 // Returns the canonicalized string for the given URL string for the
  36 // GURLTest.Types test.
  37 std::string TypesTestCase(const char* src) {
  38   GURL gurl(src);
  39   return gurl.possibly_invalid_spec();
  40 }
  41
  42 }  // namespace
  43
  44 // Different types of URLs should be handled differently by url_util, and
  45 // handed off to different canonicalizers.
  46 TEST(GURLTest, Types) {
  47   // URLs with unknown schemes should be treated as path URLs, even when they
  48   // have things like "://".
  49   EXPECT_EQ("something:///HOSTNAME.com/",
  50             TypesTestCase("something:///HOSTNAME.com/"));
  51
  52   // In the reverse, known schemes should always trigger standard URL handling.
  53   EXPECT_EQ("http://hostname.com/", TypesTestCase("http:HOSTNAME.com"));
  54   EXPECT_EQ("http://hostname.com/", TypesTestCase("http:/HOSTNAME.com"));
  55   EXPECT_EQ("http://hostname.com/", TypesTestCase("http://HOSTNAME.com"));
  56   EXPECT_EQ("http://hostname.com/", TypesTestCase("http:///HOSTNAME.com"));
  57
  58 #ifdef WIN32
  59   // URLs that look like absolute Windows drive specs.
  60   EXPECT_EQ("file:///C:/foo.txt", TypesTestCase("c:\\foo.txt"));
  61   EXPECT_EQ("file:///Z:/foo.txt", TypesTestCase("Z|foo.txt"));
  62   EXPECT_EQ("file://server/foo.txt", TypesTestCase("\\\\server\\foo.txt"));
  63   EXPECT_EQ("file://server/foo.txt", TypesTestCase("//server/foo.txt"));
  64 #endif
  65 }
  66
  67 // Test the basic creation and querying of components in a GURL. We assume
  68 // the parser is already tested and works, so we are mostly interested if the
  69 // object does the right thing with the results.
  70 TEST(GURLTest, Components) {
  71   GURL url(WStringToUTF16(L"http://user:pass@google.com:99/foo;bar?q=a#ref"));
  72   EXPECT_TRUE(url.is_valid());
  73   EXPECT_TRUE(url.SchemeIs("http"));
  74   EXPECT_FALSE(url.SchemeIsFile());
  75
  76   // This is the narrow version of the URL, which should match the wide input.
  77   EXPECT_EQ("http://user:pass@google.com:99/foo;bar?q=a#ref", url.spec());
  78
  79   EXPECT_EQ("http", url.scheme());
  80   EXPECT_EQ("user", url.username());
  81   EXPECT_EQ("pass", url.password());
  82   EXPECT_EQ("google.com", url.host());
  83   EXPECT_EQ("99", url.port());
  84   EXPECT_EQ(99, url.IntPort());
  85   EXPECT_EQ("/foo;bar", url.path());
  86   EXPECT_EQ("q=a", url.query());
  87   EXPECT_EQ("ref", url.ref());
  88 }
  89
  90 TEST(GURLTest, Empty) {
  91   GURL url;
  92   EXPECT_FALSE(url.is_valid());
  93   EXPECT_EQ("", url.spec());
  94
  95   EXPECT_EQ("", url.scheme());
  96   EXPECT_EQ("", url.username());
  97   EXPECT_EQ("", url.password());
  98   EXPECT_EQ("", url.host());
  99   EXPECT_EQ("", url.port());
 100   EXPECT_EQ(url_parse::PORT_UNSPECIFIED, url.IntPort());
 101   EXPECT_EQ("", url.path());
 102   EXPECT_EQ("", url.query());
 103   EXPECT_EQ("", url.ref());
 104 }
 105
 106 TEST(GURLTest, Copy) {
 107   GURL url(WStringToUTF16(L"http://user:pass@google.com:99/foo;bar?q=a#ref"));
 108
 109   GURL url2(url);
 110   EXPECT_TRUE(url2.is_valid());
 111
 112   EXPECT_EQ("http://user:pass@google.com:99/foo;bar?q=a#ref", url2.spec());
 113   EXPECT_EQ("http", url2.scheme());
 114   EXPECT_EQ("user", url2.username());
 115   EXPECT_EQ("pass", url2.password());
 116   EXPECT_EQ("google.com", url2.host());
 117   EXPECT_EQ("99", url2.port());
 118   EXPECT_EQ(99, url2.IntPort());
 119   EXPECT_EQ("/foo;bar", url2.path());
 120   EXPECT_EQ("q=a", url2.query());
 121   EXPECT_EQ("ref", url2.ref());
 122
 123   // Copying of invalid URL should be invalid
 124   GURL invalid;
 125   GURL invalid2(invalid);
 126   EXPECT_FALSE(invalid2.is_valid());
 127   EXPECT_EQ("", invalid2.spec());
 128   EXPECT_EQ("", invalid2.scheme());
 129   EXPECT_EQ("", invalid2.username());
 130   EXPECT_EQ("", invalid2.password());
 131   EXPECT_EQ("", invalid2.host());
 132   EXPECT_EQ("", invalid2.port());
 133   EXPECT_EQ(url_parse::PORT_UNSPECIFIED, invalid2.IntPort());
 134   EXPECT_EQ("", invalid2.path());
 135   EXPECT_EQ("", invalid2.query());
 136   EXPECT_EQ("", invalid2.ref());
 137 }
 138
 139 TEST(GURLTest, CopyFileSystem) {
 140   GURL url(WStringToUTF16(L"filesystem:https://user:pass@google.com:99/t/foo;bar?q=a#ref"));
 141
 142   GURL url2(url);
 143   EXPECT_TRUE(url2.is_valid());
 144
 145   EXPECT_EQ("filesystem:https://user:pass@google.com:99/t/foo;bar?q=a#ref", url2.spec());
 146   EXPECT_EQ("filesystem", url2.scheme());
 147   EXPECT_EQ("", url2.username());
 148   EXPECT_EQ("", url2.password());
 149   EXPECT_EQ("", url2.host());
 150   EXPECT_EQ("", url2.port());
 151   EXPECT_EQ(url_parse::PORT_UNSPECIFIED, url2.IntPort());
 152   EXPECT_EQ("/foo;bar", url2.path());
 153   EXPECT_EQ("q=a", url2.query());
 154   EXPECT_EQ("ref", url2.ref());
 155
 156   const GURL* inner = url2.inner_url();
 157   ASSERT_TRUE(inner);
 158   EXPECT_EQ("https", inner->scheme());
 159   EXPECT_EQ("user", inner->username());
 160   EXPECT_EQ("pass", inner->password());
 161   EXPECT_EQ("google.com", inner->host());
 162   EXPECT_EQ("99", inner->port());
 163   EXPECT_EQ(99, inner->IntPort());
 164   EXPECT_EQ("/t", inner->path());
 165   EXPECT_EQ("", inner->query());
 166   EXPECT_EQ("", inner->ref());
 167 }
 168
 169 // Given an invalid URL, we should still get most of the components.
 170 TEST(GURLTest, Invalid) {
 171   GURL url("http:google.com:foo");
 172   EXPECT_FALSE(url.is_valid());
 173   EXPECT_EQ("http://google.com:foo/", url.possibly_invalid_spec());
 174
 175   EXPECT_EQ("http", url.scheme());
 176   EXPECT_EQ("", url.username());
 177   EXPECT_EQ("", url.password());
 178   EXPECT_EQ("google.com", url.host());
 179   EXPECT_EQ("foo", url.port());
 180   EXPECT_EQ(url_parse::PORT_INVALID, url.IntPort());
 181   EXPECT_EQ("/", url.path());
 182   EXPECT_EQ("", url.query());
 183   EXPECT_EQ("", url.ref());
 184 }
 185
 186 TEST(GURLTest, Resolve) {
 187   // The tricky cases for relative URL resolving are tested in the
 188   // canonicalizer unit test. Here, we just test that the GURL integration
 189   // works properly.
 190   struct ResolveCase {
 191     const char* base;
 192     const char* relative;
 193     bool expected_valid;
 194     const char* expected;
 195   } resolve_cases[] = {
 196     {"http://www.google.com/", "foo.html", true, "http://www.google.com/foo.html"},
 197     {"http://www.google.com/", "http://images.google.com/foo.html", true, "http://images.google.com/foo.html"},
 198     {"http://www.google.com/blah/bloo?c#d", "../../../hello/./world.html?a#b", true, "http://www.google.com/hello/world.html?a#b"},
 199     {"http://www.google.com/foo#bar", "#com", true, "http://www.google.com/foo#com"},
 200     {"http://www.google.com/", "Https:images.google.com", true, "https://images.google.com/"},
 201       // A non-standard base can be replaced with a standard absolute URL.
 202     {"data:blahblah", "http://google.com/", true, "http://google.com/"},
 203     {"data:blahblah", "http:google.com", true, "http://google.com/"},
 204       // Filesystem URLs have different paths to test.
 205     {"filesystem:http://www.google.com/type/", "foo.html", true, "filesystem:http://www.google.com/type/foo.html"},
 206     {"filesystem:http://www.google.com/type/", "../foo.html", true, "filesystem:http://www.google.com/type/foo.html"},
 207   };
 208
 209   for (size_t i = 0; i < ARRAYSIZE(resolve_cases); i++) {
 210     // 8-bit code path.
 211     GURL input(resolve_cases[i].base);
 212     GURL output = input.Resolve(resolve_cases[i].relative);
 213     EXPECT_EQ(resolve_cases[i].expected_valid, output.is_valid()) << i;
 214     EXPECT_EQ(resolve_cases[i].expected, output.spec()) << i;
 215     EXPECT_EQ(output.SchemeIsFileSystem(), output.inner_url() != NULL);
 216
 217     // Wide code path.
 218     GURL inputw(ConvertUTF8ToUTF16(resolve_cases[i].base));
 219     GURL outputw =
 220         input.Resolve(ConvertUTF8ToUTF16(resolve_cases[i].relative));
 221     EXPECT_EQ(resolve_cases[i].expected_valid, outputw.is_valid()) << i;
 222     EXPECT_EQ(resolve_cases[i].expected, outputw.spec()) << i;
 223     EXPECT_EQ(outputw.SchemeIsFileSystem(), outputw.inner_url() != NULL);
 224   }
 225 }
 226
 227 TEST(GURLTest, GetOrigin) {
 228   struct TestCase {
 229     const char* input;
 230     const char* expected;
 231   } cases[] = {
 232     {"http://www.google.com", "http://www.google.com/"},
 233     {"javascript:window.alert(\"hello,world\");", ""},
 234     {"http://user:pass@www.google.com:21/blah#baz", "http://www.google.com:21/"},
 235     {"http://user@www.google.com", "http://www.google.com/"},
 236     {"http://:pass@www.google.com", "http://www.google.com/"},
 237     {"http://:@www.google.com", "http://www.google.com/"},
 238     {"filesystem:http://www.google.com/temp/foo?q#b", "http://www.google.com/"},
 239     {"filesystem:http://user:pass@google.com:21/blah#baz", "http://google.com:21/"},
 240   };
 241   for (size_t i = 0; i < ARRAYSIZE(cases); i++) {
 242     GURL url(cases[i].input);
 243     GURL origin = url.GetOrigin();
 244     EXPECT_EQ(cases[i].expected, origin.spec());
 245   }
 246 }
 247
 248 TEST(GURLTest, GetWithEmptyPath) {
 249   struct TestCase {
 250     const char* input;
 251     const char* expected;
 252   } cases[] = {
 253     {"http://www.google.com", "http://www.google.com/"},
 254     {"javascript:window.alert(\"hello, world\");", ""},
 255     {"http://www.google.com/foo/bar.html?baz=22", "http://www.google.com/"},
 256     {"filesystem:http://www.google.com/temporary/bar.html?baz=22", "filesystem:http://www.google.com/temporary/"},
 257     {"filesystem:file:///temporary/bar.html?baz=22", "filesystem:file:///temporary/"},
 258   };
 259
 260   for (size_t i = 0; i < ARRAYSIZE(cases); i++) {
 261     GURL url(cases[i].input);
 262     GURL empty_path = url.GetWithEmptyPath();
 263     EXPECT_EQ(cases[i].expected, empty_path.spec());
 264   }
 265 }
 266
 267 TEST(GURLTest, Replacements) {
 268   // The url canonicalizer replacement test will handle most of these case.
 269   // The most important thing to do here is to check that the proper
 270   // canonicalizer gets called based on the scheme of the input.
 271   struct ReplaceCase {
 272     const char* base;
 273     const char* scheme;
 274     const char* username;
 275     const char* password;
 276     const char* host;
 277     const char* port;
 278     const char* path;
 279     const char* query;
 280     const char* ref;
 281     const char* expected;
 282   } replace_cases[] = {
 283     {"http://www.google.com/foo/bar.html?foo#bar", NULL, NULL, NULL, NULL, NULL, "/", "", "", "http://www.google.com/"},
 284     {"http://www.google.com/foo/bar.html?foo#bar", "javascript", "", "", "", "", "window.open('foo');", "", "", "javascript:window.open('foo');"},
 285     {"file:///C:/foo/bar.txt", "http", NULL, NULL, "www.google.com", "99", "/foo","search", "ref", "http://www.google.com:99/foo?search#ref"},
 286 #ifdef WIN32
 287     {"http://www.google.com/foo/bar.html?foo#bar", "file", "", "", "", "", "c:\\", "", "", "file:///C:/"},
 288 #endif
 289     {"filesystem:http://www.google.com/foo/bar.html?foo#bar", NULL, NULL, NULL, NULL, NULL, "/", "", "", "filesystem:http://www.google.com/foo/"},
 290   };
 291
 292   for (size_t i = 0; i < ARRAYSIZE(replace_cases); i++) {
 293     const ReplaceCase& cur = replace_cases[i];
 294     GURL url(cur.base);
 295     GURL::Replacements repl;
 296     SetupReplacement(&GURL::Replacements::SetScheme, &repl, cur.scheme);
 297     SetupReplacement(&GURL::Replacements::SetUsername, &repl, cur.username);
 298     SetupReplacement(&GURL::Replacements::SetPassword, &repl, cur.password);
 299     SetupReplacement(&GURL::Replacements::SetHost, &repl, cur.host);
 300     SetupReplacement(&GURL::Replacements::SetPort, &repl, cur.port);
 301     SetupReplacement(&GURL::Replacements::SetPath, &repl, cur.path);
 302     SetupReplacement(&GURL::Replacements::SetQuery, &repl, cur.query);
 303     SetupReplacement(&GURL::Replacements::SetRef, &repl, cur.ref);
 304     GURL output = url.ReplaceComponents(repl);
 305
 306     EXPECT_EQ(replace_cases[i].expected, output.spec());
 307     EXPECT_EQ(output.SchemeIsFileSystem(), output.inner_url() != NULL);
 308   }
 309 }
 310
 311 TEST(GURLTest, PathForRequest) {
 312   struct TestCase {
 313     const char* input;
 314     const char* expected;
 315     const char* inner_expected;
 316   } cases[] = {
 317     {"http://www.google.com", "/", NULL},
 318     {"http://www.google.com/", "/", NULL},
 319     {"http://www.google.com/foo/bar.html?baz=22", "/foo/bar.html?baz=22", NULL},
 320     {"http://www.google.com/foo/bar.html#ref", "/foo/bar.html", NULL},
 321     {"http://www.google.com/foo/bar.html?query#ref", "/foo/bar.html?query", NULL},
 322     {"filesystem:http://www.google.com/temporary/foo/bar.html?query#ref", "/foo/bar.html?query", "/temporary"},
 323     {"filesystem:http://www.google.com/temporary/foo/bar.html?query", "/foo/bar.html?query", "/temporary"},
 324   };
 325
 326   for (size_t i = 0; i < ARRAYSIZE(cases); i++) {
 327     GURL url(cases[i].input);
 328     std::string path_request = url.PathForRequest();
 329     EXPECT_EQ(cases[i].expected, path_request);
 330     EXPECT_EQ(cases[i].inner_expected == NULL, url.inner_url() == NULL);
 331     if (url.inner_url() && cases[i].inner_expected)
 332       EXPECT_EQ(cases[i].inner_expected, url.inner_url()->PathForRequest());
 333   }
 334 }
 335
 336 TEST(GURLTest, EffectiveIntPort) {
 337   struct PortTest {
 338     const char* spec;
 339     int expected_int_port;
 340   } port_tests[] = {
 341     // http
 342     {"http://www.google.com/", 80},
 343     {"http://www.google.com:80/", 80},
 344     {"http://www.google.com:443/", 443},
 345
 346     // https
 347     {"https://www.google.com/", 443},
 348     {"https://www.google.com:443/", 443},
 349     {"https://www.google.com:80/", 80},
 350
 351     // ftp
 352     {"ftp://www.google.com/", 21},
 353     {"ftp://www.google.com:21/", 21},
 354     {"ftp://www.google.com:80/", 80},
 355
 356     // gopher
 357     {"gopher://www.google.com/", 70},
 358     {"gopher://www.google.com:70/", 70},
 359     {"gopher://www.google.com:80/", 80},
 360
 361     // file - no port
 362     {"file://www.google.com/", url_parse::PORT_UNSPECIFIED},
 363     {"file://www.google.com:443/", url_parse::PORT_UNSPECIFIED},
 364
 365     // data - no port
 366     {"data:www.google.com:90", url_parse::PORT_UNSPECIFIED},
 367     {"data:www.google.com", url_parse::PORT_UNSPECIFIED},
 368
 369     // filesystem - no port
 370     {"filesystem:http://www.google.com:90/t/foo", url_parse::PORT_UNSPECIFIED},
 371     {"filesystem:file:///t/foo", url_parse::PORT_UNSPECIFIED},
 372   };
 373
 374   for (size_t i = 0; i < ARRAYSIZE(port_tests); i++) {
 375     GURL url(port_tests[i].spec);
 376     EXPECT_EQ(port_tests[i].expected_int_port, url.EffectiveIntPort());
 377   }
 378 }
 379
 380 TEST(GURLTest, IPAddress) {
 381   struct IPTest {
 382     const char* spec;
 383     bool expected_ip;
 384   } ip_tests[] = {
 385     {"http://www.google.com/", false},
 386     {"http://192.168.9.1/", true},
 387     {"http://192.168.9.1.2/", false},
 388     {"http://192.168.m.1/", false},
 389     {"http://2001:db8::1/", false},
 390     {"http://[2001:db8::1]/", true},
 391     {"", false},
 392     {"some random input!", false},
 393   };
 394
 395   for (size_t i = 0; i < ARRAYSIZE(ip_tests); i++) {
 396     GURL url(ip_tests[i].spec);
 397     EXPECT_EQ(ip_tests[i].expected_ip, url.HostIsIPAddress());
 398   }
 399 }
 400
 401 TEST(GURLTest, HostNoBrackets) {
 402   struct TestCase {
 403     const char* input;
 404     const char* expected_host;
 405     const char* expected_plainhost;
 406   } cases[] = {
 407     {"http://www.google.com", "www.google.com", "www.google.com"},
 408     {"http://[2001:db8::1]/", "[2001:db8::1]", "2001:db8::1"},
 409     {"http://[::]/", "[::]", "::"},
 410
 411     // Don't require a valid URL, but don't crash either.
 412     {"http://[]/", "[]", ""},
 413     {"http://[x]/", "[x]", "x"},
 414     {"http://[x/", "[x", "[x"},
 415     {"http://x]/", "x]", "x]"},
 416     {"http://[/", "[", "["},
 417     {"http://]/", "]", "]"},
 418     {"", "", ""},
 419   };
 420   for (size_t i = 0; i < ARRAYSIZE(cases); i++) {
 421     GURL url(cases[i].input);
 422     EXPECT_EQ(cases[i].expected_host, url.host());
 423     EXPECT_EQ(cases[i].expected_plainhost, url.HostNoBrackets());
 424   }
 425 }
 426
 427 TEST(GURLTest, DomainIs) {
 428   const char google_domain[] = "google.com";
 429
 430   GURL url_1("http://www.google.com:99/foo");
 431   EXPECT_TRUE(url_1.DomainIs(google_domain));
 432
 433   GURL url_2("http://google.com:99/foo");
 434   EXPECT_TRUE(url_2.DomainIs(google_domain));
 435
 436   GURL url_3("http://google.com./foo");
 437   EXPECT_TRUE(url_3.DomainIs(google_domain));
 438
 439   GURL url_4("http://google.com/foo");
 440   EXPECT_FALSE(url_4.DomainIs("google.com."));
 441
 442   GURL url_5("http://google.com./foo");
 443   EXPECT_TRUE(url_5.DomainIs("google.com."));
 444
 445   GURL url_6("http://www.google.com./foo");
 446   EXPECT_TRUE(url_6.DomainIs(".com."));
 447
 448   GURL url_7("http://www.balabala.com/foo");
 449   EXPECT_FALSE(url_7.DomainIs(google_domain));
 450
 451   GURL url_8("http://www.google.com.cn/foo");
 452   EXPECT_FALSE(url_8.DomainIs(google_domain));
 453
 454   GURL url_9("http://www.iamnotgoogle.com/foo");
 455   EXPECT_FALSE(url_9.DomainIs(google_domain));
 456
 457   GURL url_10("http://www.iamnotgoogle.com../foo");
 458   EXPECT_FALSE(url_10.DomainIs(".com"));
 459
 460   GURL url_11("filesystem:http://www.google.com:99/foo/");
 461   EXPECT_TRUE(url_11.DomainIs(google_domain));
 462
 463   GURL url_12("filesystem:http://www.iamnotgoogle.com/foo/");
 464   EXPECT_FALSE(url_12.DomainIs(google_domain));
 465 }
 466
 467 // Newlines should be stripped from inputs.
 468 TEST(GURLTest, Newlines) {
 469   // Constructor.
 470   GURL url_1(" \t ht\ntp://\twww.goo\rgle.com/as\ndf \n ");
 471   EXPECT_EQ("http://www.google.com/asdf", url_1.spec());
 472
 473   // Relative path resolver.
 474   GURL url_2 = url_1.Resolve(" \n /fo\to\r ");
 475   EXPECT_EQ("http://www.google.com/foo", url_2.spec());
 476
 477   // Note that newlines are NOT stripped from ReplaceComponents.
 478 }
 479
 480 TEST(GURLTest, IsStandard) {
 481   GURL a("http:foo/bar");
 482   EXPECT_TRUE(a.IsStandard());
 483
 484   GURL b("foo:bar/baz");
 485   EXPECT_FALSE(b.IsStandard());
 486
 487   GURL c("foo://bar/baz");
 488   EXPECT_FALSE(c.IsStandard());
 489 }