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