punycode_test.go 4.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161
  1. // Copyright 2012 The Go Authors. All rights reserved.
  2. // Use of this source code is governed by a BSD-style
  3. // license that can be found in the LICENSE file.
  4. package cookiejar
  5. import (
  6. "testing"
  7. )
  8. var punycodeTestCases = [...]struct {
  9. s, encoded string
  10. }{
  11. {"", ""},
  12. {"-", "--"},
  13. {"-a", "-a-"},
  14. {"-a-", "-a--"},
  15. {"a", "a-"},
  16. {"a-", "a--"},
  17. {"a-b", "a-b-"},
  18. {"books", "books-"},
  19. {"bücher", "bcher-kva"},
  20. {"Hello世界", "Hello-ck1hg65u"},
  21. {"ü", "tda"},
  22. {"üý", "tdac"},
  23. // The test cases below come from RFC 3492 section 7.1 with Errata 3026.
  24. {
  25. // (A) Arabic (Egyptian).
  26. "\u0644\u064A\u0647\u0645\u0627\u0628\u062A\u0643\u0644" +
  27. "\u0645\u0648\u0634\u0639\u0631\u0628\u064A\u061F",
  28. "egbpdaj6bu4bxfgehfvwxn",
  29. },
  30. {
  31. // (B) Chinese (simplified).
  32. "\u4ED6\u4EEC\u4E3A\u4EC0\u4E48\u4E0D\u8BF4\u4E2D\u6587",
  33. "ihqwcrb4cv8a8dqg056pqjye",
  34. },
  35. {
  36. // (C) Chinese (traditional).
  37. "\u4ED6\u5011\u7232\u4EC0\u9EBD\u4E0D\u8AAA\u4E2D\u6587",
  38. "ihqwctvzc91f659drss3x8bo0yb",
  39. },
  40. {
  41. // (D) Czech.
  42. "\u0050\u0072\u006F\u010D\u0070\u0072\u006F\u0073\u0074" +
  43. "\u011B\u006E\u0065\u006D\u006C\u0075\u0076\u00ED\u010D" +
  44. "\u0065\u0073\u006B\u0079",
  45. "Proprostnemluvesky-uyb24dma41a",
  46. },
  47. {
  48. // (E) Hebrew.
  49. "\u05DC\u05DE\u05D4\u05D4\u05DD\u05E4\u05E9\u05D5\u05D8" +
  50. "\u05DC\u05D0\u05DE\u05D3\u05D1\u05E8\u05D9\u05DD\u05E2" +
  51. "\u05D1\u05E8\u05D9\u05EA",
  52. "4dbcagdahymbxekheh6e0a7fei0b",
  53. },
  54. {
  55. // (F) Hindi (Devanagari).
  56. "\u092F\u0939\u0932\u094B\u0917\u0939\u093F\u0928\u094D" +
  57. "\u0926\u0940\u0915\u094D\u092F\u094B\u0902\u0928\u0939" +
  58. "\u0940\u0902\u092C\u094B\u0932\u0938\u0915\u0924\u0947" +
  59. "\u0939\u0948\u0902",
  60. "i1baa7eci9glrd9b2ae1bj0hfcgg6iyaf8o0a1dig0cd",
  61. },
  62. {
  63. // (G) Japanese (kanji and hiragana).
  64. "\u306A\u305C\u307F\u3093\u306A\u65E5\u672C\u8A9E\u3092" +
  65. "\u8A71\u3057\u3066\u304F\u308C\u306A\u3044\u306E\u304B",
  66. "n8jok5ay5dzabd5bym9f0cm5685rrjetr6pdxa",
  67. },
  68. {
  69. // (H) Korean (Hangul syllables).
  70. "\uC138\uACC4\uC758\uBAA8\uB4E0\uC0AC\uB78C\uB4E4\uC774" +
  71. "\uD55C\uAD6D\uC5B4\uB97C\uC774\uD574\uD55C\uB2E4\uBA74" +
  72. "\uC5BC\uB9C8\uB098\uC88B\uC744\uAE4C",
  73. "989aomsvi5e83db1d2a355cv1e0vak1dwrv93d5xbh15a0dt30a5j" +
  74. "psd879ccm6fea98c",
  75. },
  76. {
  77. // (I) Russian (Cyrillic).
  78. "\u043F\u043E\u0447\u0435\u043C\u0443\u0436\u0435\u043E" +
  79. "\u043D\u0438\u043D\u0435\u0433\u043E\u0432\u043E\u0440" +
  80. "\u044F\u0442\u043F\u043E\u0440\u0443\u0441\u0441\u043A" +
  81. "\u0438",
  82. "b1abfaaepdrnnbgefbadotcwatmq2g4l",
  83. },
  84. {
  85. // (J) Spanish.
  86. "\u0050\u006F\u0072\u0071\u0075\u00E9\u006E\u006F\u0070" +
  87. "\u0075\u0065\u0064\u0065\u006E\u0073\u0069\u006D\u0070" +
  88. "\u006C\u0065\u006D\u0065\u006E\u0074\u0065\u0068\u0061" +
  89. "\u0062\u006C\u0061\u0072\u0065\u006E\u0045\u0073\u0070" +
  90. "\u0061\u00F1\u006F\u006C",
  91. "PorqunopuedensimplementehablarenEspaol-fmd56a",
  92. },
  93. {
  94. // (K) Vietnamese.
  95. "\u0054\u1EA1\u0069\u0073\u0061\u006F\u0068\u1ECD\u006B" +
  96. "\u0068\u00F4\u006E\u0067\u0074\u0068\u1EC3\u0063\u0068" +
  97. "\u1EC9\u006E\u00F3\u0069\u0074\u0069\u1EBF\u006E\u0067" +
  98. "\u0056\u0069\u1EC7\u0074",
  99. "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g",
  100. },
  101. {
  102. // (L) 3<nen>B<gumi><kinpachi><sensei>.
  103. "\u0033\u5E74\u0042\u7D44\u91D1\u516B\u5148\u751F",
  104. "3B-ww4c5e180e575a65lsy2b",
  105. },
  106. {
  107. // (M) <amuro><namie>-with-SUPER-MONKEYS.
  108. "\u5B89\u5BA4\u5948\u7F8E\u6075\u002D\u0077\u0069\u0074" +
  109. "\u0068\u002D\u0053\u0055\u0050\u0045\u0052\u002D\u004D" +
  110. "\u004F\u004E\u004B\u0045\u0059\u0053",
  111. "-with-SUPER-MONKEYS-pc58ag80a8qai00g7n9n",
  112. },
  113. {
  114. // (N) Hello-Another-Way-<sorezore><no><basho>.
  115. "\u0048\u0065\u006C\u006C\u006F\u002D\u0041\u006E\u006F" +
  116. "\u0074\u0068\u0065\u0072\u002D\u0057\u0061\u0079\u002D" +
  117. "\u305D\u308C\u305E\u308C\u306E\u5834\u6240",
  118. "Hello-Another-Way--fc4qua05auwb3674vfr0b",
  119. },
  120. {
  121. // (O) <hitotsu><yane><no><shita>2.
  122. "\u3072\u3068\u3064\u5C4B\u6839\u306E\u4E0B\u0032",
  123. "2-u9tlzr9756bt3uc0v",
  124. },
  125. {
  126. // (P) Maji<de>Koi<suru>5<byou><mae>
  127. "\u004D\u0061\u006A\u0069\u3067\u004B\u006F\u0069\u3059" +
  128. "\u308B\u0035\u79D2\u524D",
  129. "MajiKoi5-783gue6qz075azm5e",
  130. },
  131. {
  132. // (Q) <pafii>de<runba>
  133. "\u30D1\u30D5\u30A3\u30FC\u0064\u0065\u30EB\u30F3\u30D0",
  134. "de-jg4avhby1noc0d",
  135. },
  136. {
  137. // (R) <sono><supiido><de>
  138. "\u305D\u306E\u30B9\u30D4\u30FC\u30C9\u3067",
  139. "d9juau41awczczp",
  140. },
  141. {
  142. // (S) -> $1.00 <-
  143. "\u002D\u003E\u0020\u0024\u0031\u002E\u0030\u0030\u0020" +
  144. "\u003C\u002D",
  145. "-> $1.00 <--",
  146. },
  147. }
  148. func TestPunycode(t *testing.T) {
  149. for _, tc := range punycodeTestCases {
  150. if got, err := encode("", tc.s); err != nil {
  151. t.Errorf(`encode("", %q): %v`, tc.s, err)
  152. } else if got != tc.encoded {
  153. t.Errorf(`encode("", %q): got %q, want %q`, tc.s, got, tc.encoded)
  154. }
  155. }
  156. }