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