XmlChar.cs 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374
  1. // -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*-
  2. //
  3. // System.Xml.XmlChar.cs
  4. //
  5. // Author:
  6. // Jason Diamond ([email protected])
  7. //
  8. // (C) 2001 Jason Diamond http://injektilo.org/
  9. //
  10. //
  11. // Permission is hereby granted, free of charge, to any person obtaining
  12. // a copy of this software and associated documentation files (the
  13. // "Software"), to deal in the Software without restriction, including
  14. // without limitation the rights to use, copy, modify, merge, publish,
  15. // distribute, sublicense, and/or sell copies of the Software, and to
  16. // permit persons to whom the Software is furnished to do so, subject to
  17. // the following conditions:
  18. //
  19. // The above copyright notice and this permission notice shall be
  20. // included in all copies or substantial portions of the Software.
  21. //
  22. // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  23. // EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  24. // MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  25. // NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
  26. // LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
  27. // OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
  28. // WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  29. //
  30. namespace System.Xml
  31. {
  32. internal class XmlChar
  33. {
  34. public static readonly char [] WhitespaceChars = new char [] {' ', '\n', '\t', '\r'};
  35. public static bool IsWhitespace (int ch)
  36. {
  37. return ch == 0x20 || ch == 0x9 || ch == 0xD || ch == 0xA;
  38. }
  39. public static bool IsWhitespace (string str)
  40. {
  41. for (int i = 0; i < str.Length; i++)
  42. if (!IsWhitespace (str [i])) return false;
  43. return true;
  44. }
  45. public static bool IsFirstNameChar (int ch)
  46. {
  47. bool result = false;
  48. if (ch >= 0 && ch <= 0xFFFF)
  49. {
  50. result = (nameBitmap[(firstNamePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  51. }
  52. return result;
  53. }
  54. public static bool IsValid (int ch)
  55. {
  56. return !IsInvalid (ch);
  57. }
  58. public static bool IsInvalid (int ch)
  59. {
  60. switch (ch) {
  61. case 9:
  62. case 10:
  63. case 13:
  64. return false;
  65. }
  66. if (ch < 32)
  67. return true;
  68. if (ch < 0xD800)
  69. return false;
  70. if (ch < 0xE000)
  71. return true;
  72. if (ch < 0xFFFE)
  73. return false;
  74. if (ch < 0x10000)
  75. return true;
  76. if (ch < 0x110000)
  77. return false;
  78. else
  79. return true;
  80. }
  81. public static bool IsNameChar (int ch)
  82. {
  83. bool result = false;
  84. if (ch >= 0 && ch <= 0xFFFF)
  85. {
  86. result = (nameBitmap[(namePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  87. }
  88. return result;
  89. }
  90. public static bool IsNCNameChar (int ch)
  91. {
  92. bool result = false;
  93. if (ch >= 0 && ch <= 0xFFFF && ch != ':')
  94. {
  95. result = (nameBitmap[(namePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  96. }
  97. return result;
  98. }
  99. public static bool IsName (string str)
  100. {
  101. if (str.Length == 0)
  102. return false;
  103. if (!IsFirstNameChar (str [0]))
  104. return false;
  105. for (int i = 1; i < str.Length; i++)
  106. if (!IsNameChar (str [i]))
  107. return false;
  108. return true;
  109. }
  110. public static bool IsNCName (string str)
  111. {
  112. if (str.Length == 0)
  113. return false;
  114. if (!IsFirstNameChar (str [0]))
  115. return false;
  116. for (int i = 0; i < str.Length; i++)
  117. if (!IsNCNameChar (str [i]))
  118. return false;
  119. return true;
  120. }
  121. public static bool IsNmToken (string str)
  122. {
  123. if (str.Length == 0)
  124. return false;
  125. for (int i = 0; i < str.Length; i++)
  126. if (!IsNameChar (str [i]))
  127. return false;
  128. return true;
  129. }
  130. public static bool IsPubidChar (int ch)
  131. {
  132. return (IsWhitespace(ch) && ch != '\t') | ('a' <= ch && ch <= 'z') | ('A' <= ch && ch <= 'Z') | ('0' <= ch && ch <= '9') | "-'()+,./:=?;!*#@$_%".IndexOf((char)ch) >= 0;
  133. }
  134. public static bool IsPubid (string str)
  135. {
  136. for (int i = 0; i < str.Length; i++)
  137. if (!IsPubidChar (str [i]))
  138. return false;
  139. return true;
  140. }
  141. // encodings (copied from XmlConstructs.cs)
  142. /// <summary>
  143. /// Returns true if the encoding name is a valid IANA encoding.
  144. /// This method does not verify that there is a decoder available
  145. /// for this encoding, only that the characters are valid for an
  146. /// IANA encoding name.
  147. /// </summary>
  148. /// <param name="ianaEncoding">The encoding to check.</param>
  149. /// <returns></returns>
  150. public static bool IsValidIANAEncoding (String ianaEncoding)
  151. {
  152. if (ianaEncoding != null)
  153. {
  154. int length = ianaEncoding.Length;
  155. if (length > 0)
  156. {
  157. char c = ianaEncoding[0];
  158. if ((c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z'))
  159. {
  160. for (int i = 1; i < length; i++)
  161. {
  162. c = ianaEncoding[i];
  163. if ((c < 'A' || c > 'Z') && (c < 'a' || c > 'z') &&
  164. (c < '0' || c > '9') && c != '.' && c != '_' &&
  165. c != '-')
  166. {
  167. return false;
  168. }
  169. }
  170. return true;
  171. }
  172. }
  173. }
  174. return false;
  175. }
  176. public static int GetPredefinedEntity (string name)
  177. {
  178. switch (name) {
  179. case "amp":
  180. return '&';
  181. case "lt":
  182. return '<';
  183. case "gt":
  184. return '>';
  185. case "quot":
  186. return '"';
  187. case "apos":
  188. return '\'';
  189. default:
  190. return -1;
  191. }
  192. }
  193. static readonly byte [] firstNamePages =
  194. {
  195. 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x00,
  196. 0x00, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
  197. 0x10, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  198. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x13,
  199. 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  200. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  201. 0x15, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  202. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  203. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  204. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01,
  205. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  206. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  207. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  208. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  209. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  210. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  211. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  212. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  213. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  214. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x17,
  215. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  216. 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01,
  217. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  218. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  219. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  220. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  221. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x18,
  222. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  223. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  224. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  225. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  226. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
  227. };
  228. static readonly byte [] namePages =
  229. {
  230. 0x19, 0x03, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x00,
  231. 0x00, 0x1F, 0x20, 0x21, 0x22, 0x23, 0x24, 0x25,
  232. 0x10, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  233. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x13,
  234. 0x26, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  235. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  236. 0x27, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  237. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  238. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  239. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01,
  240. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  241. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  242. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  243. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  244. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  245. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  246. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  247. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  248. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  249. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x17,
  250. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  251. 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01,
  252. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  253. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  254. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  255. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  256. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x18,
  257. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  258. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  259. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  260. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  261. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
  262. };
  263. static readonly uint [] nameBitmap =
  264. {
  265. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  266. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  267. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  268. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  269. 0x00000000, 0x04000000, 0x87FFFFFE, 0x07FFFFFE,
  270. 0x00000000, 0x00000000, 0xFF7FFFFF, 0xFF7FFFFF,
  271. 0xFFFFFFFF, 0x7FF3FFFF, 0xFFFFFDFE, 0x7FFFFFFF,
  272. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFE00F, 0xFC31FFFF,
  273. 0x00FFFFFF, 0x00000000, 0xFFFF0000, 0xFFFFFFFF,
  274. 0xFFFFFFFF, 0xF80001FF, 0x00000003, 0x00000000,
  275. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  276. 0xFFFFD740, 0xFFFFFFFB, 0x547F7FFF, 0x000FFFFD,
  277. 0xFFFFDFFE, 0xFFFFFFFF, 0xDFFEFFFF, 0xFFFFFFFF,
  278. 0xFFFF0003, 0xFFFFFFFF, 0xFFFF199F, 0x033FCFFF,
  279. 0x00000000, 0xFFFE0000, 0x027FFFFF, 0xFFFFFFFE,
  280. 0x0000007F, 0x00000000, 0xFFFF0000, 0x000707FF,
  281. 0x00000000, 0x07FFFFFE, 0x000007FE, 0xFFFE0000,
  282. 0xFFFFFFFF, 0x7CFFFFFF, 0x002F7FFF, 0x00000060,
  283. 0xFFFFFFE0, 0x23FFFFFF, 0xFF000000, 0x00000003,
  284. 0xFFF99FE0, 0x03C5FDFF, 0xB0000000, 0x00030003,
  285. 0xFFF987E0, 0x036DFDFF, 0x5E000000, 0x001C0000,
  286. 0xFFFBAFE0, 0x23EDFDFF, 0x00000000, 0x00000001,
  287. 0xFFF99FE0, 0x23CDFDFF, 0xB0000000, 0x00000003,
  288. 0xD63DC7E0, 0x03BFC718, 0x00000000, 0x00000000,
  289. 0xFFFDDFE0, 0x03EFFDFF, 0x00000000, 0x00000003,
  290. 0xFFFDDFE0, 0x03EFFDFF, 0x40000000, 0x00000003,
  291. 0xFFFDDFE0, 0x03FFFDFF, 0x00000000, 0x00000003,
  292. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  293. 0xFFFFFFFE, 0x000D7FFF, 0x0000003F, 0x00000000,
  294. 0xFEF02596, 0x200D6CAE, 0x0000001F, 0x00000000,
  295. 0x00000000, 0x00000000, 0xFFFFFEFF, 0x000003FF,
  296. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  297. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  298. 0x00000000, 0xFFFFFFFF, 0xFFFF003F, 0x007FFFFF,
  299. 0x0007DAED, 0x50000000, 0x82315001, 0x002C62AB,
  300. 0x40000000, 0xF580C900, 0x00000007, 0x02010800,
  301. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  302. 0x0FFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x03FFFFFF,
  303. 0x3F3FFFFF, 0xFFFFFFFF, 0xAAFF3F3F, 0x3FFFFFFF,
  304. 0xFFFFFFFF, 0x5FDFFFFF, 0x0FCF1FDC, 0x1FDC1FFF,
  305. 0x00000000, 0x00004C40, 0x00000000, 0x00000000,
  306. 0x00000007, 0x00000000, 0x00000000, 0x00000000,
  307. 0x00000080, 0x000003FE, 0xFFFFFFFE, 0xFFFFFFFF,
  308. 0x001FFFFF, 0xFFFFFFFE, 0xFFFFFFFF, 0x07FFFFFF,
  309. 0xFFFFFFE0, 0x00001FFF, 0x00000000, 0x00000000,
  310. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  311. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  312. 0xFFFFFFFF, 0x0000003F, 0x00000000, 0x00000000,
  313. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  314. 0xFFFFFFFF, 0x0000000F, 0x00000000, 0x00000000,
  315. 0x00000000, 0x07FF6000, 0x87FFFFFE, 0x07FFFFFE,
  316. 0x00000000, 0x00800000, 0xFF7FFFFF, 0xFF7FFFFF,
  317. 0x00FFFFFF, 0x00000000, 0xFFFF0000, 0xFFFFFFFF,
  318. 0xFFFFFFFF, 0xF80001FF, 0x00030003, 0x00000000,
  319. 0xFFFFFFFF, 0xFFFFFFFF, 0x0000003F, 0x00000003,
  320. 0xFFFFD7C0, 0xFFFFFFFB, 0x547F7FFF, 0x000FFFFD,
  321. 0xFFFFDFFE, 0xFFFFFFFF, 0xDFFEFFFF, 0xFFFFFFFF,
  322. 0xFFFF007B, 0xFFFFFFFF, 0xFFFF199F, 0x033FCFFF,
  323. 0x00000000, 0xFFFE0000, 0x027FFFFF, 0xFFFFFFFE,
  324. 0xFFFE007F, 0xBBFFFFFB, 0xFFFF0016, 0x000707FF,
  325. 0x00000000, 0x07FFFFFE, 0x0007FFFF, 0xFFFF03FF,
  326. 0xFFFFFFFF, 0x7CFFFFFF, 0xFFEF7FFF, 0x03FF3DFF,
  327. 0xFFFFFFEE, 0xF3FFFFFF, 0xFF1E3FFF, 0x0000FFCF,
  328. 0xFFF99FEE, 0xD3C5FDFF, 0xB080399F, 0x0003FFCF,
  329. 0xFFF987E4, 0xD36DFDFF, 0x5E003987, 0x001FFFC0,
  330. 0xFFFBAFEE, 0xF3EDFDFF, 0x00003BBF, 0x0000FFC1,
  331. 0xFFF99FEE, 0xF3CDFDFF, 0xB0C0398F, 0x0000FFC3,
  332. 0xD63DC7EC, 0xC3BFC718, 0x00803DC7, 0x0000FF80,
  333. 0xFFFDDFEE, 0xC3EFFDFF, 0x00603DDF, 0x0000FFC3,
  334. 0xFFFDDFEC, 0xC3EFFDFF, 0x40603DDF, 0x0000FFC3,
  335. 0xFFFDDFEC, 0xC3FFFDFF, 0x00803DCF, 0x0000FFC3,
  336. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  337. 0xFFFFFFFE, 0x07FF7FFF, 0x03FF7FFF, 0x00000000,
  338. 0xFEF02596, 0x3BFF6CAE, 0x03FF3F5F, 0x00000000,
  339. 0x03000000, 0xC2A003FF, 0xFFFFFEFF, 0xFFFE03FF,
  340. 0xFEBF0FDF, 0x02FE3FFF, 0x00000000, 0x00000000,
  341. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  342. 0x00000000, 0x00000000, 0x1FFF0000, 0x00000002,
  343. 0x000000A0, 0x003EFFFE, 0xFFFFFFFE, 0xFFFFFFFF,
  344. 0x661FFFFF, 0xFFFFFFFE, 0xFFFFFFFF, 0x77FFFFFF
  345. };
  346. }
  347. }