url/url_util_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 "base/macros.h"
   6 #include "testing/gtest/include/gtest/gtest.h"
   7 #include "url/url_canon.h"
   8 #include "url/url_canon_stdstring.h"
   9 #include "url/url_parse.h"
  10 #include "url/url_test_utils.h"
  11 #include "url/url_util.h"
  12
  13 namespace url {
  14
  15 TEST(URLUtilTest, FindAndCompareScheme) {
  16   Component found_scheme;
  17
  18   // Simple case where the scheme is found and matches.
  19   const char kStr1[] = "http://www.com/";
  20   EXPECT_TRUE(FindAndCompareScheme(
  21       kStr1, static_cast<int>(strlen(kStr1)), "http", NULL));
  22   EXPECT_TRUE(FindAndCompareScheme(
  23       kStr1, static_cast<int>(strlen(kStr1)), "http", &found_scheme));
  24   EXPECT_TRUE(found_scheme == Component(0, 4));
  25
  26   // A case where the scheme is found and doesn't match.
  27   EXPECT_FALSE(FindAndCompareScheme(
  28       kStr1, static_cast<int>(strlen(kStr1)), "https", &found_scheme));
  29   EXPECT_TRUE(found_scheme == Component(0, 4));
  30
  31   // A case where there is no scheme.
  32   const char kStr2[] = "httpfoobar";
  33   EXPECT_FALSE(FindAndCompareScheme(
  34       kStr2, static_cast<int>(strlen(kStr2)), "http", &found_scheme));
  35   EXPECT_TRUE(found_scheme == Component());
  36
  37   // When there is an empty scheme, it should match the empty scheme.
  38   const char kStr3[] = ":foo.com/";
  39   EXPECT_TRUE(FindAndCompareScheme(
  40       kStr3, static_cast<int>(strlen(kStr3)), "", &found_scheme));
  41   EXPECT_TRUE(found_scheme == Component(0, 0));
  42
  43   // But when there is no scheme, it should fail.
  44   EXPECT_FALSE(FindAndCompareScheme("", 0, "", &found_scheme));
  45   EXPECT_TRUE(found_scheme == Component());
  46
  47   // When there is a whitespace char in scheme, it should canonicalize the url
  48   // before comparison.
  49   const char whtspc_str[] = " \r\n\tjav\ra\nscri\tpt:alert(1)";
  50   EXPECT_TRUE(FindAndCompareScheme(whtspc_str,
  51                                    static_cast<int>(strlen(whtspc_str)),
  52                                    "javascript", &found_scheme));
  53   EXPECT_TRUE(found_scheme == Component(1, 10));
  54
  55   // Control characters should be stripped out on the ends, and kept in the
  56   // middle.
  57   const char ctrl_str[] = "\02jav\02scr\03ipt:alert(1)";
  58   EXPECT_FALSE(FindAndCompareScheme(ctrl_str,
  59                                     static_cast<int>(strlen(ctrl_str)),
  60                                     "javascript", &found_scheme));
  61   EXPECT_TRUE(found_scheme == Component(1, 11));
  62 }
  63
  64 TEST(URLUtilTest, ReplaceComponents) {
  65   Parsed parsed;
  66   RawCanonOutputT<char> output;
  67   Parsed new_parsed;
  68
  69   // Check that the following calls do not cause crash
  70   Replacements<char> replacements;
  71   replacements.SetRef("test", Component(0, 4));
  72   ReplaceComponents(NULL, 0, parsed, replacements, NULL, &output, &new_parsed);
  73   ReplaceComponents("", 0, parsed, replacements, NULL, &output, &new_parsed);
  74   replacements.ClearRef();
  75   replacements.SetHost("test", Component(0, 4));
  76   ReplaceComponents(NULL, 0, parsed, replacements, NULL, &output, &new_parsed);
  77   ReplaceComponents("", 0, parsed, replacements, NULL, &output, &new_parsed);
  78
  79   replacements.ClearHost();
  80   ReplaceComponents(NULL, 0, parsed, replacements, NULL, &output, &new_parsed);
  81   ReplaceComponents("", 0, parsed, replacements, NULL, &output, &new_parsed);
  82   ReplaceComponents(NULL, 0, parsed, replacements, NULL, &output, &new_parsed);
  83   ReplaceComponents("", 0, parsed, replacements, NULL, &output, &new_parsed);
  84 }
  85
  86 static std::string CheckReplaceScheme(const char* base_url,
  87                                       const char* scheme) {
  88   // Make sure the input is canonicalized.
  89   RawCanonOutput<32> original;
  90   Parsed original_parsed;
  91   Canonicalize(base_url, strlen(base_url), true, NULL, &original,
  92                &original_parsed);
  93
  94   Replacements<char> replacements;
  95   replacements.SetScheme(scheme, Component(0, strlen(scheme)));
  96
  97   std::string output_string;
  98   StdStringCanonOutput output(&output_string);
  99   Parsed output_parsed;
 100   ReplaceComponents(original.data(), original.length(), original_parsed,
 101                     replacements, NULL, &output, &output_parsed);
 102
 103   output.Complete();
 104   return output_string;
 105 }
 106
 107 TEST(URLUtilTest, ReplaceScheme) {
 108   EXPECT_EQ("https://google.com/",
 109             CheckReplaceScheme("http://google.com/", "https"));
 110   EXPECT_EQ("file://google.com/",
 111             CheckReplaceScheme("http://google.com/", "file"));
 112   EXPECT_EQ("http://home/Build",
 113             CheckReplaceScheme("file:///Home/Build", "http"));
 114   EXPECT_EQ("javascript:foo",
 115             CheckReplaceScheme("about:foo", "javascript"));
 116   EXPECT_EQ("://google.com/",
 117             CheckReplaceScheme("http://google.com/", ""));
 118   EXPECT_EQ("http://google.com/",
 119             CheckReplaceScheme("about:google.com", "http"));
 120   EXPECT_EQ("http:", CheckReplaceScheme("", "http"));
 121
 122 #ifdef WIN32
 123   // Magic Windows drive letter behavior when converting to a file URL.
 124   EXPECT_EQ("file:///E:/foo/",
 125             CheckReplaceScheme("http://localhost/e:foo/", "file"));
 126 #endif
 127
 128   // This will probably change to "about://google.com/" when we fix
 129   // http://crbug.com/160 which should also be an acceptable result.
 130   EXPECT_EQ("about://google.com/",
 131             CheckReplaceScheme("http://google.com/", "about"));
 132
 133   EXPECT_EQ("http://example.com/%20hello%20# world",
 134             CheckReplaceScheme("myscheme:example.com/ hello # world ", "http"));
 135 }
 136
 137 TEST(URLUtilTest, DecodeURLEscapeSequences) {
 138   struct DecodeCase {
 139     const char* input;
 140     const char* output;
 141   } decode_cases[] = {
 142     {"hello, world", "hello, world"},
 143     {"%01%02%03%04%05%06%07%08%09%0a%0B%0C%0D%0e%0f/",
 144      "\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0B\x0C\x0D\x0e\x0f/"},
 145     {"%10%11%12%13%14%15%16%17%18%19%1a%1B%1C%1D%1e%1f/",
 146      "\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1B\x1C\x1D\x1e\x1f/"},
 147     {"%20%21%22%23%24%25%26%27%28%29%2a%2B%2C%2D%2e%2f/",
 148      " !\"#$%&'()*+,-.//"},
 149     {"%30%31%32%33%34%35%36%37%38%39%3a%3B%3C%3D%3e%3f/",
 150      "0123456789:;<=>?/"},
 151     {"%40%41%42%43%44%45%46%47%48%49%4a%4B%4C%4D%4e%4f/",
 152      "@ABCDEFGHIJKLMNO/"},
 153     {"%50%51%52%53%54%55%56%57%58%59%5a%5B%5C%5D%5e%5f/",
 154      "PQRSTUVWXYZ[\\]^_/"},
 155     {"%60%61%62%63%64%65%66%67%68%69%6a%6B%6C%6D%6e%6f/",
 156      "`abcdefghijklmno/"},
 157     {"%70%71%72%73%74%75%76%77%78%79%7a%7B%7C%7D%7e%7f/",
 158      "pqrstuvwxyz{|}~\x7f/"},
 159     // Test un-UTF-8-ization.
 160     {"%e4%bd%a0%e5%a5%bd", "\xe4\xbd\xa0\xe5\xa5\xbd"},
 161   };
 162
 163   for (size_t i = 0; i < arraysize(decode_cases); i++) {
 164     const char* input = decode_cases[i].input;
 165     RawCanonOutputT<base::char16> output;
 166     DecodeURLEscapeSequences(input, strlen(input), &output);
 167     EXPECT_EQ(decode_cases[i].output,
 168               test_utils::ConvertUTF16ToUTF8(base::string16(output.data(),
 169                                                             output.length())));
 170   }
 171
 172   // Our decode should decode %00
 173   const char zero_input[] = "%00";
 174   RawCanonOutputT<base::char16> zero_output;
 175   DecodeURLEscapeSequences(zero_input, strlen(zero_input), &zero_output);
 176   EXPECT_NE("%00", test_utils::ConvertUTF16ToUTF8(
 177       base::string16(zero_output.data(), zero_output.length())));
 178
 179   // Test the error behavior for invalid UTF-8.
 180   const char invalid_input[] = "%e4%a0%e5%a5%bd";
 181   const base::char16 invalid_expected[4] = {0x00e4, 0x00a0, 0x597d, 0};
 182   RawCanonOutputT<base::char16> invalid_output;
 183   DecodeURLEscapeSequences(invalid_input, strlen(invalid_input),
 184                            &invalid_output);
 185   EXPECT_EQ(base::string16(invalid_expected),
 186             base::string16(invalid_output.data(), invalid_output.length()));
 187 }
 188
 189 TEST(URLUtilTest, TestEncodeURIComponent) {
 190   struct EncodeCase {
 191     const char* input;
 192     const char* output;
 193   } encode_cases[] = {
 194     {"hello, world", "hello%2C%20world"},
 195     {"\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F",
 196      "%01%02%03%04%05%06%07%08%09%0A%0B%0C%0D%0E%0F"},
 197     {"\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1A\x1B\x1C\x1D\x1E\x1F",
 198      "%10%11%12%13%14%15%16%17%18%19%1A%1B%1C%1D%1E%1F"},
 199     {" !\"#$%&'()*+,-./",
 200      "%20!%22%23%24%25%26%27()*%2B%2C-.%2F"},
 201     {"0123456789:;<=>?",
 202      "0123456789%3A%3B%3C%3D%3E%3F"},
 203     {"@ABCDEFGHIJKLMNO",
 204      "%40ABCDEFGHIJKLMNO"},
 205     {"PQRSTUVWXYZ[\\]^_",
 206      "PQRSTUVWXYZ%5B%5C%5D%5E_"},
 207     {"`abcdefghijklmno",
 208      "%60abcdefghijklmno"},
 209     {"pqrstuvwxyz{|}~\x7f",
 210      "pqrstuvwxyz%7B%7C%7D~%7F"},
 211   };
 212
 213   for (size_t i = 0; i < arraysize(encode_cases); i++) {
 214     const char* input = encode_cases[i].input;
 215     RawCanonOutputT<char> buffer;
 216     EncodeURIComponent(input, strlen(input), &buffer);
 217     std::string output(buffer.data(), buffer.length());
 218     EXPECT_EQ(encode_cases[i].output, output);
 219   }
 220 }
 221
 222 TEST(URLUtilTest, TestResolveRelativeWithNonStandardBase) {
 223   // This tests non-standard (in the sense that GIsStandard() == false)
 224   // hierarchical schemes.
 225   struct ResolveRelativeCase {
 226     const char* base;
 227     const char* rel;
 228     bool is_valid;
 229     const char* out;
 230   } resolve_non_standard_cases[] = {
 231       // Resolving a relative path against a non-hierarchical URL should fail.
 232     {"scheme:opaque_data", "/path", false, ""},
 233       // Resolving a relative path against a non-standard authority-based base
 234       // URL doesn't alter the authority section.
 235     {"scheme://Authority/", "../path", true, "scheme://Authority/path"},
 236       // A non-standard hierarchical base is resolved with path URL
 237       // canonicalization rules.
 238     {"data:/Blah:Blah/", "file.html", true, "data:/Blah:Blah/file.html"},
 239     {"data:/Path/../part/part2", "file.html", true,
 240       "data:/Path/../part/file.html"},
 241       // Path URL canonicalization rules also apply to non-standard authority-
 242       // based URLs.
 243     {"custom://Authority/", "file.html", true,
 244       "custom://Authority/file.html"},
 245     {"custom://Authority/", "other://Auth/", true, "other://Auth/"},
 246     {"custom://Authority/", "../../file.html", true,
 247       "custom://Authority/file.html"},
 248     {"custom://Authority/path/", "file.html", true,
 249       "custom://Authority/path/file.html"},
 250     {"custom://Authority:NoCanon/path/", "file.html", true,
 251       "custom://Authority:NoCanon/path/file.html"},
 252       // It's still possible to get an invalid path URL.
 253     {"custom://Invalid:!#Auth/", "file.html", false, ""},
 254       // A path with an authority section gets canonicalized under standard URL
 255       // rules, even though the base was non-standard.
 256     {"content://content.Provider/", "//other.Provider", true,
 257       "content://other.provider/"},
 258       // Resolving an absolute URL doesn't cause canonicalization of the
 259       // result.
 260     {"about:blank", "custom://Authority", true, "custom://Authority"},
 261       // Fragment URLs can be resolved against a non-standard base.
 262     {"scheme://Authority/path", "#fragment", true,
 263       "scheme://Authority/path#fragment"},
 264     {"scheme://Authority/", "#fragment", true, "scheme://Authority/#fragment"},
 265       // Resolving should fail if the base URL is authority-based but is
 266       // missing a path component (the '/' at the end).
 267     {"scheme://Authority", "path", false, ""},
 268       // Test resolving a fragment (only) against any kind of base-URL.
 269     {"about:blank", "#id42", true, "about:blank#id42" },
 270     {"about:blank", " #id42", true, "about:blank#id42" },
 271     {"about:blank#oldfrag", "#newfrag", true, "about:blank#newfrag" },
 272       // A surprising side effect of allowing fragments to resolve against
 273       // any URL scheme is we might break javascript: URLs by doing so...
 274     {"javascript:alert('foo#bar')", "#badfrag", true,
 275       "javascript:alert('foo#badfrag" },
 276       // In this case, the backslashes will not be canonicalized because it's a
 277       // non-standard URL, but they will be treated as a path separators,
 278       // giving the base URL here a path of "\".
 279       //
 280       // The result here is somewhat arbitrary. One could argue it should be
 281       // either "aaa://a\" or "aaa://a/" since the path is being replaced with
 282       // the "current directory". But in the context of resolving on data URLs,
 283       // adding the requested dot doesn't seem wrong either.
 284     {"aaa://a\\", "aaa:.", true, "aaa://a\\." }
 285   };
 286
 287   for (size_t i = 0; i < arraysize(resolve_non_standard_cases); i++) {
 288     const ResolveRelativeCase& test_data = resolve_non_standard_cases[i];
 289     Parsed base_parsed;
 290     ParsePathURL(test_data.base, strlen(test_data.base), false, &base_parsed);
 291
 292     std::string resolved;
 293     StdStringCanonOutput output(&resolved);
 294     Parsed resolved_parsed;
 295     bool valid = ResolveRelative(test_data.base, strlen(test_data.base),
 296                                  base_parsed, test_data.rel,
 297                                  strlen(test_data.rel), NULL, &output,
 298                                  &resolved_parsed);
 299     output.Complete();
 300
 301     EXPECT_EQ(test_data.is_valid, valid) << i;
 302     if (test_data.is_valid && valid)
 303       EXPECT_EQ(test_data.out, resolved) << i;
 304   }
 305 }
 306
 307 TEST(URLUtilTest, TestNoRefComponent) {
 308   // The hash-mark must be ignored when mailto: scheme is
 309   // parsed, even if the url has a base and relative part.
 310   const char* base = "mailto://to/";
 311   const char* rel = "any#body";
 312
 313   Parsed base_parsed;
 314   ParsePathURL(base, strlen(base), false, &base_parsed);
 315
 316   std::string resolved;
 317   StdStringCanonOutput output(&resolved);
 318   Parsed resolved_parsed;
 319
 320   bool valid = ResolveRelative(base, strlen(base),
 321                                base_parsed, rel,
 322                                strlen(rel), NULL, &output,
 323                                &resolved_parsed);
 324   EXPECT_TRUE(valid);
 325   EXPECT_FALSE(resolved_parsed.ref.is_valid());
 326 }
 327
 328 }  // namespace url