XmlChar.cs 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379
  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. // Now, "XmlChar" and "XmlConstructs" are made as equivalent, so
  33. // I dicided to rename XmlConstruts class as "XmlChar" and use it
  34. // for default build.
  35. // However, this class will be used for the future compact framework
  36. // (XmlConstruts class uses not a little memory).
  37. internal class XmlCharCompact
  38. {
  39. public static readonly char [] WhitespaceChars = new char [] {' ', '\n', '\t', '\r'};
  40. public static bool IsWhitespace (int ch)
  41. {
  42. return ch == 0x20 || ch == 0x9 || ch == 0xD || ch == 0xA;
  43. }
  44. public static bool IsWhitespace (string str)
  45. {
  46. for (int i = 0; i < str.Length; i++)
  47. if (!IsWhitespace (str [i])) return false;
  48. return true;
  49. }
  50. public static bool IsFirstNameChar (int ch)
  51. {
  52. bool result = false;
  53. if (ch >= 0 && ch <= 0xFFFF)
  54. {
  55. result = (nameBitmap[(firstNamePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  56. }
  57. return result;
  58. }
  59. public static bool IsValid (int ch)
  60. {
  61. return !IsInvalid (ch);
  62. }
  63. public static bool IsInvalid (int ch)
  64. {
  65. switch (ch) {
  66. case 9:
  67. case 10:
  68. case 13:
  69. return false;
  70. }
  71. if (ch < 32)
  72. return true;
  73. if (ch < 0xD800)
  74. return false;
  75. if (ch < 0xE000)
  76. return true;
  77. if (ch < 0xFFFE)
  78. return false;
  79. if (ch < 0x10000)
  80. return true;
  81. if (ch < 0x110000)
  82. return false;
  83. else
  84. return true;
  85. }
  86. public static bool IsNameChar (int ch)
  87. {
  88. bool result = false;
  89. if (ch >= 0 && ch <= 0xFFFF)
  90. {
  91. result = (nameBitmap[(namePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  92. }
  93. return result;
  94. }
  95. public static bool IsNCNameChar (int ch)
  96. {
  97. bool result = false;
  98. if (ch >= 0 && ch <= 0xFFFF && ch != ':')
  99. {
  100. result = (nameBitmap[(namePages[ch >> 8] << 3) + ((ch & 0xFF) >> 5)] & (1 << (ch & 0x1F))) != 0;
  101. }
  102. return result;
  103. }
  104. public static bool IsName (string str)
  105. {
  106. if (str.Length == 0)
  107. return false;
  108. if (!IsFirstNameChar (str [0]))
  109. return false;
  110. for (int i = 1; i < str.Length; i++)
  111. if (!IsNameChar (str [i]))
  112. return false;
  113. return true;
  114. }
  115. public static bool IsNCName (string str)
  116. {
  117. if (str.Length == 0)
  118. return false;
  119. if (!IsFirstNameChar (str [0]))
  120. return false;
  121. for (int i = 0; i < str.Length; i++)
  122. if (!IsNCNameChar (str [i]))
  123. return false;
  124. return true;
  125. }
  126. public static bool IsNmToken (string str)
  127. {
  128. if (str.Length == 0)
  129. return false;
  130. for (int i = 0; i < str.Length; i++)
  131. if (!IsNameChar (str [i]))
  132. return false;
  133. return true;
  134. }
  135. public static bool IsPubidChar (int ch)
  136. {
  137. return (IsWhitespace(ch) && ch != '\t') | ('a' <= ch && ch <= 'z') | ('A' <= ch && ch <= 'Z') | ('0' <= ch && ch <= '9') | "-'()+,./:=?;!*#@$_%".IndexOf((char)ch) >= 0;
  138. }
  139. public static bool IsPubid (string str)
  140. {
  141. for (int i = 0; i < str.Length; i++)
  142. if (!IsPubidChar (str [i]))
  143. return false;
  144. return true;
  145. }
  146. // encodings (copied from XmlConstructs.cs)
  147. /// <summary>
  148. /// Returns true if the encoding name is a valid IANA encoding.
  149. /// This method does not verify that there is a decoder available
  150. /// for this encoding, only that the characters are valid for an
  151. /// IANA encoding name.
  152. /// </summary>
  153. /// <param name="ianaEncoding">The encoding to check.</param>
  154. /// <returns></returns>
  155. public static bool IsValidIANAEncoding (String ianaEncoding)
  156. {
  157. if (ianaEncoding != null)
  158. {
  159. int length = ianaEncoding.Length;
  160. if (length > 0)
  161. {
  162. char c = ianaEncoding[0];
  163. if ((c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z'))
  164. {
  165. for (int i = 1; i < length; i++)
  166. {
  167. c = ianaEncoding[i];
  168. if ((c < 'A' || c > 'Z') && (c < 'a' || c > 'z') &&
  169. (c < '0' || c > '9') && c != '.' && c != '_' &&
  170. c != '-')
  171. {
  172. return false;
  173. }
  174. }
  175. return true;
  176. }
  177. }
  178. }
  179. return false;
  180. }
  181. public static int GetPredefinedEntity (string name)
  182. {
  183. switch (name) {
  184. case "amp":
  185. return '&';
  186. case "lt":
  187. return '<';
  188. case "gt":
  189. return '>';
  190. case "quot":
  191. return '"';
  192. case "apos":
  193. return '\'';
  194. default:
  195. return -1;
  196. }
  197. }
  198. static readonly byte [] firstNamePages =
  199. {
  200. 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x00,
  201. 0x00, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
  202. 0x10, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  203. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x13,
  204. 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  205. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  206. 0x15, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  207. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  208. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  209. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 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, 0x01,
  215. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  216. 0x01, 0x01, 0x01, 0x01, 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, 0x17,
  220. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  221. 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01,
  222. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  223. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  224. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  225. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  226. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x18,
  227. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  228. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  229. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  230. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  231. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
  232. };
  233. static readonly byte [] namePages =
  234. {
  235. 0x19, 0x03, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x00,
  236. 0x00, 0x1F, 0x20, 0x21, 0x22, 0x23, 0x24, 0x25,
  237. 0x10, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  238. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x13,
  239. 0x26, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  240. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  241. 0x27, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  242. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  243. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  244. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 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, 0x01,
  250. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  251. 0x01, 0x01, 0x01, 0x01, 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, 0x17,
  255. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  256. 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01,
  257. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  258. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  259. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  260. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  261. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x18,
  262. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  263. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  264. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  265. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  266. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
  267. };
  268. static readonly uint [] nameBitmap =
  269. {
  270. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  271. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  272. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  273. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  274. 0x00000000, 0x04000000, 0x87FFFFFE, 0x07FFFFFE,
  275. 0x00000000, 0x00000000, 0xFF7FFFFF, 0xFF7FFFFF,
  276. 0xFFFFFFFF, 0x7FF3FFFF, 0xFFFFFDFE, 0x7FFFFFFF,
  277. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFE00F, 0xFC31FFFF,
  278. 0x00FFFFFF, 0x00000000, 0xFFFF0000, 0xFFFFFFFF,
  279. 0xFFFFFFFF, 0xF80001FF, 0x00000003, 0x00000000,
  280. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  281. 0xFFFFD740, 0xFFFFFFFB, 0x547F7FFF, 0x000FFFFD,
  282. 0xFFFFDFFE, 0xFFFFFFFF, 0xDFFEFFFF, 0xFFFFFFFF,
  283. 0xFFFF0003, 0xFFFFFFFF, 0xFFFF199F, 0x033FCFFF,
  284. 0x00000000, 0xFFFE0000, 0x027FFFFF, 0xFFFFFFFE,
  285. 0x0000007F, 0x00000000, 0xFFFF0000, 0x000707FF,
  286. 0x00000000, 0x07FFFFFE, 0x000007FE, 0xFFFE0000,
  287. 0xFFFFFFFF, 0x7CFFFFFF, 0x002F7FFF, 0x00000060,
  288. 0xFFFFFFE0, 0x23FFFFFF, 0xFF000000, 0x00000003,
  289. 0xFFF99FE0, 0x03C5FDFF, 0xB0000000, 0x00030003,
  290. 0xFFF987E0, 0x036DFDFF, 0x5E000000, 0x001C0000,
  291. 0xFFFBAFE0, 0x23EDFDFF, 0x00000000, 0x00000001,
  292. 0xFFF99FE0, 0x23CDFDFF, 0xB0000000, 0x00000003,
  293. 0xD63DC7E0, 0x03BFC718, 0x00000000, 0x00000000,
  294. 0xFFFDDFE0, 0x03EFFDFF, 0x00000000, 0x00000003,
  295. 0xFFFDDFE0, 0x03EFFDFF, 0x40000000, 0x00000003,
  296. 0xFFFDDFE0, 0x03FFFDFF, 0x00000000, 0x00000003,
  297. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  298. 0xFFFFFFFE, 0x000D7FFF, 0x0000003F, 0x00000000,
  299. 0xFEF02596, 0x200D6CAE, 0x0000001F, 0x00000000,
  300. 0x00000000, 0x00000000, 0xFFFFFEFF, 0x000003FF,
  301. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  302. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  303. 0x00000000, 0xFFFFFFFF, 0xFFFF003F, 0x007FFFFF,
  304. 0x0007DAED, 0x50000000, 0x82315001, 0x002C62AB,
  305. 0x40000000, 0xF580C900, 0x00000007, 0x02010800,
  306. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  307. 0x0FFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x03FFFFFF,
  308. 0x3F3FFFFF, 0xFFFFFFFF, 0xAAFF3F3F, 0x3FFFFFFF,
  309. 0xFFFFFFFF, 0x5FDFFFFF, 0x0FCF1FDC, 0x1FDC1FFF,
  310. 0x00000000, 0x00004C40, 0x00000000, 0x00000000,
  311. 0x00000007, 0x00000000, 0x00000000, 0x00000000,
  312. 0x00000080, 0x000003FE, 0xFFFFFFFE, 0xFFFFFFFF,
  313. 0x001FFFFF, 0xFFFFFFFE, 0xFFFFFFFF, 0x07FFFFFF,
  314. 0xFFFFFFE0, 0x00001FFF, 0x00000000, 0x00000000,
  315. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  316. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  317. 0xFFFFFFFF, 0x0000003F, 0x00000000, 0x00000000,
  318. 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
  319. 0xFFFFFFFF, 0x0000000F, 0x00000000, 0x00000000,
  320. 0x00000000, 0x07FF6000, 0x87FFFFFE, 0x07FFFFFE,
  321. 0x00000000, 0x00800000, 0xFF7FFFFF, 0xFF7FFFFF,
  322. 0x00FFFFFF, 0x00000000, 0xFFFF0000, 0xFFFFFFFF,
  323. 0xFFFFFFFF, 0xF80001FF, 0x00030003, 0x00000000,
  324. 0xFFFFFFFF, 0xFFFFFFFF, 0x0000003F, 0x00000003,
  325. 0xFFFFD7C0, 0xFFFFFFFB, 0x547F7FFF, 0x000FFFFD,
  326. 0xFFFFDFFE, 0xFFFFFFFF, 0xDFFEFFFF, 0xFFFFFFFF,
  327. 0xFFFF007B, 0xFFFFFFFF, 0xFFFF199F, 0x033FCFFF,
  328. 0x00000000, 0xFFFE0000, 0x027FFFFF, 0xFFFFFFFE,
  329. 0xFFFE007F, 0xBBFFFFFB, 0xFFFF0016, 0x000707FF,
  330. 0x00000000, 0x07FFFFFE, 0x0007FFFF, 0xFFFF03FF,
  331. 0xFFFFFFFF, 0x7CFFFFFF, 0xFFEF7FFF, 0x03FF3DFF,
  332. 0xFFFFFFEE, 0xF3FFFFFF, 0xFF1E3FFF, 0x0000FFCF,
  333. 0xFFF99FEE, 0xD3C5FDFF, 0xB080399F, 0x0003FFCF,
  334. 0xFFF987E4, 0xD36DFDFF, 0x5E003987, 0x001FFFC0,
  335. 0xFFFBAFEE, 0xF3EDFDFF, 0x00003BBF, 0x0000FFC1,
  336. 0xFFF99FEE, 0xF3CDFDFF, 0xB0C0398F, 0x0000FFC3,
  337. 0xD63DC7EC, 0xC3BFC718, 0x00803DC7, 0x0000FF80,
  338. 0xFFFDDFEE, 0xC3EFFDFF, 0x00603DDF, 0x0000FFC3,
  339. 0xFFFDDFEC, 0xC3EFFDFF, 0x40603DDF, 0x0000FFC3,
  340. 0xFFFDDFEC, 0xC3FFFDFF, 0x00803DCF, 0x0000FFC3,
  341. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  342. 0xFFFFFFFE, 0x07FF7FFF, 0x03FF7FFF, 0x00000000,
  343. 0xFEF02596, 0x3BFF6CAE, 0x03FF3F5F, 0x00000000,
  344. 0x03000000, 0xC2A003FF, 0xFFFFFEFF, 0xFFFE03FF,
  345. 0xFEBF0FDF, 0x02FE3FFF, 0x00000000, 0x00000000,
  346. 0x00000000, 0x00000000, 0x00000000, 0x00000000,
  347. 0x00000000, 0x00000000, 0x1FFF0000, 0x00000002,
  348. 0x000000A0, 0x003EFFFE, 0xFFFFFFFE, 0xFFFFFFFF,
  349. 0x661FFFFF, 0xFFFFFFFE, 0xFFFFFFFF, 0x77FFFFFF
  350. };
  351. }
  352. }