chvalid.c 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334
  1. /*
  2. * chvalid.c: this module implements the character range
  3. * validation APIs
  4. *
  5. * This file is automatically generated from the cvs source
  6. * definition files using the genChRanges.py Python script
  7. *
  8. * Generation date: Mon Mar 27 11:09:48 2006
  9. * Sources: chvalid.def
  10. * William Brack <wbrack@mmm.com.hk>
  11. */
  12. #define IN_LIBXML
  13. #include "libxml.h"
  14. #include <libxml/chvalid.h>
  15. /*
  16. * The initial tables ({func_name}_tab) are used to validate whether a
  17. * single-byte character is within the specified group. Each table
  18. * contains 256 bytes, with each byte representing one of the 256
  19. * possible characters. If the table byte is set, the character is
  20. * allowed.
  21. *
  22. */
  23. const unsigned char xmlIsPubidChar_tab[256] = {
  24. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00,
  25. 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  26. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x00, 0x01,
  27. 0x01, 0x01, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  28. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  29. 0x00, 0x01, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  30. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  31. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x01,
  32. 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  33. 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
  34. 0x01, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  35. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  36. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  37. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  38. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  39. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  40. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  41. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  42. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  43. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  44. 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  45. 0x00, 0x00, 0x00, 0x00 };
  46. static const xmlChSRange xmlIsBaseChar_srng[] = { {0x100, 0x131},
  47. {0x134, 0x13e}, {0x141, 0x148}, {0x14a, 0x17e}, {0x180, 0x1c3},
  48. {0x1cd, 0x1f0}, {0x1f4, 0x1f5}, {0x1fa, 0x217}, {0x250, 0x2a8},
  49. {0x2bb, 0x2c1}, {0x386, 0x386}, {0x388, 0x38a}, {0x38c, 0x38c},
  50. {0x38e, 0x3a1}, {0x3a3, 0x3ce}, {0x3d0, 0x3d6}, {0x3da, 0x3da},
  51. {0x3dc, 0x3dc}, {0x3de, 0x3de}, {0x3e0, 0x3e0}, {0x3e2, 0x3f3},
  52. {0x401, 0x40c}, {0x40e, 0x44f}, {0x451, 0x45c}, {0x45e, 0x481},
  53. {0x490, 0x4c4}, {0x4c7, 0x4c8}, {0x4cb, 0x4cc}, {0x4d0, 0x4eb},
  54. {0x4ee, 0x4f5}, {0x4f8, 0x4f9}, {0x531, 0x556}, {0x559, 0x559},
  55. {0x561, 0x586}, {0x5d0, 0x5ea}, {0x5f0, 0x5f2}, {0x621, 0x63a},
  56. {0x641, 0x64a}, {0x671, 0x6b7}, {0x6ba, 0x6be}, {0x6c0, 0x6ce},
  57. {0x6d0, 0x6d3}, {0x6d5, 0x6d5}, {0x6e5, 0x6e6}, {0x905, 0x939},
  58. {0x93d, 0x93d}, {0x958, 0x961}, {0x985, 0x98c}, {0x98f, 0x990},
  59. {0x993, 0x9a8}, {0x9aa, 0x9b0}, {0x9b2, 0x9b2}, {0x9b6, 0x9b9},
  60. {0x9dc, 0x9dd}, {0x9df, 0x9e1}, {0x9f0, 0x9f1}, {0xa05, 0xa0a},
  61. {0xa0f, 0xa10}, {0xa13, 0xa28}, {0xa2a, 0xa30}, {0xa32, 0xa33},
  62. {0xa35, 0xa36}, {0xa38, 0xa39}, {0xa59, 0xa5c}, {0xa5e, 0xa5e},
  63. {0xa72, 0xa74}, {0xa85, 0xa8b}, {0xa8d, 0xa8d}, {0xa8f, 0xa91},
  64. {0xa93, 0xaa8}, {0xaaa, 0xab0}, {0xab2, 0xab3}, {0xab5, 0xab9},
  65. {0xabd, 0xabd}, {0xae0, 0xae0}, {0xb05, 0xb0c}, {0xb0f, 0xb10},
  66. {0xb13, 0xb28}, {0xb2a, 0xb30}, {0xb32, 0xb33}, {0xb36, 0xb39},
  67. {0xb3d, 0xb3d}, {0xb5c, 0xb5d}, {0xb5f, 0xb61}, {0xb85, 0xb8a},
  68. {0xb8e, 0xb90}, {0xb92, 0xb95}, {0xb99, 0xb9a}, {0xb9c, 0xb9c},
  69. {0xb9e, 0xb9f}, {0xba3, 0xba4}, {0xba8, 0xbaa}, {0xbae, 0xbb5},
  70. {0xbb7, 0xbb9}, {0xc05, 0xc0c}, {0xc0e, 0xc10}, {0xc12, 0xc28},
  71. {0xc2a, 0xc33}, {0xc35, 0xc39}, {0xc60, 0xc61}, {0xc85, 0xc8c},
  72. {0xc8e, 0xc90}, {0xc92, 0xca8}, {0xcaa, 0xcb3}, {0xcb5, 0xcb9},
  73. {0xcde, 0xcde}, {0xce0, 0xce1}, {0xd05, 0xd0c}, {0xd0e, 0xd10},
  74. {0xd12, 0xd28}, {0xd2a, 0xd39}, {0xd60, 0xd61}, {0xe01, 0xe2e},
  75. {0xe30, 0xe30}, {0xe32, 0xe33}, {0xe40, 0xe45}, {0xe81, 0xe82},
  76. {0xe84, 0xe84}, {0xe87, 0xe88}, {0xe8a, 0xe8a}, {0xe8d, 0xe8d},
  77. {0xe94, 0xe97}, {0xe99, 0xe9f}, {0xea1, 0xea3}, {0xea5, 0xea5},
  78. {0xea7, 0xea7}, {0xeaa, 0xeab}, {0xead, 0xeae}, {0xeb0, 0xeb0},
  79. {0xeb2, 0xeb3}, {0xebd, 0xebd}, {0xec0, 0xec4}, {0xf40, 0xf47},
  80. {0xf49, 0xf69}, {0x10a0, 0x10c5}, {0x10d0, 0x10f6}, {0x1100, 0x1100},
  81. {0x1102, 0x1103}, {0x1105, 0x1107}, {0x1109, 0x1109}, {0x110b, 0x110c},
  82. {0x110e, 0x1112}, {0x113c, 0x113c}, {0x113e, 0x113e}, {0x1140, 0x1140},
  83. {0x114c, 0x114c}, {0x114e, 0x114e}, {0x1150, 0x1150}, {0x1154, 0x1155},
  84. {0x1159, 0x1159}, {0x115f, 0x1161}, {0x1163, 0x1163}, {0x1165, 0x1165},
  85. {0x1167, 0x1167}, {0x1169, 0x1169}, {0x116d, 0x116e}, {0x1172, 0x1173},
  86. {0x1175, 0x1175}, {0x119e, 0x119e}, {0x11a8, 0x11a8}, {0x11ab, 0x11ab},
  87. {0x11ae, 0x11af}, {0x11b7, 0x11b8}, {0x11ba, 0x11ba}, {0x11bc, 0x11c2},
  88. {0x11eb, 0x11eb}, {0x11f0, 0x11f0}, {0x11f9, 0x11f9}, {0x1e00, 0x1e9b},
  89. {0x1ea0, 0x1ef9}, {0x1f00, 0x1f15}, {0x1f18, 0x1f1d}, {0x1f20, 0x1f45},
  90. {0x1f48, 0x1f4d}, {0x1f50, 0x1f57}, {0x1f59, 0x1f59}, {0x1f5b, 0x1f5b},
  91. {0x1f5d, 0x1f5d}, {0x1f5f, 0x1f7d}, {0x1f80, 0x1fb4}, {0x1fb6, 0x1fbc},
  92. {0x1fbe, 0x1fbe}, {0x1fc2, 0x1fc4}, {0x1fc6, 0x1fcc}, {0x1fd0, 0x1fd3},
  93. {0x1fd6, 0x1fdb}, {0x1fe0, 0x1fec}, {0x1ff2, 0x1ff4}, {0x1ff6, 0x1ffc},
  94. {0x2126, 0x2126}, {0x212a, 0x212b}, {0x212e, 0x212e}, {0x2180, 0x2182},
  95. {0x3041, 0x3094}, {0x30a1, 0x30fa}, {0x3105, 0x312c}, {0xac00, 0xd7a3}};
  96. const xmlChRangeGroup xmlIsBaseCharGroup =
  97. {197, 0, xmlIsBaseChar_srng, (xmlChLRangePtr)0};
  98. static const xmlChSRange xmlIsChar_srng[] = { {0x100, 0xd7ff},
  99. {0xe000, 0xfffd}};
  100. static const xmlChLRange xmlIsChar_lrng[] = { {0x10000, 0x10ffff}};
  101. const xmlChRangeGroup xmlIsCharGroup =
  102. {2, 1, xmlIsChar_srng, xmlIsChar_lrng};
  103. static const xmlChSRange xmlIsCombining_srng[] = { {0x300, 0x345},
  104. {0x360, 0x361}, {0x483, 0x486}, {0x591, 0x5a1}, {0x5a3, 0x5b9},
  105. {0x5bb, 0x5bd}, {0x5bf, 0x5bf}, {0x5c1, 0x5c2}, {0x5c4, 0x5c4},
  106. {0x64b, 0x652}, {0x670, 0x670}, {0x6d6, 0x6dc}, {0x6dd, 0x6df},
  107. {0x6e0, 0x6e4}, {0x6e7, 0x6e8}, {0x6ea, 0x6ed}, {0x901, 0x903},
  108. {0x93c, 0x93c}, {0x93e, 0x94c}, {0x94d, 0x94d}, {0x951, 0x954},
  109. {0x962, 0x963}, {0x981, 0x983}, {0x9bc, 0x9bc}, {0x9be, 0x9be},
  110. {0x9bf, 0x9bf}, {0x9c0, 0x9c4}, {0x9c7, 0x9c8}, {0x9cb, 0x9cd},
  111. {0x9d7, 0x9d7}, {0x9e2, 0x9e3}, {0xa02, 0xa02}, {0xa3c, 0xa3c},
  112. {0xa3e, 0xa3e}, {0xa3f, 0xa3f}, {0xa40, 0xa42}, {0xa47, 0xa48},
  113. {0xa4b, 0xa4d}, {0xa70, 0xa71}, {0xa81, 0xa83}, {0xabc, 0xabc},
  114. {0xabe, 0xac5}, {0xac7, 0xac9}, {0xacb, 0xacd}, {0xb01, 0xb03},
  115. {0xb3c, 0xb3c}, {0xb3e, 0xb43}, {0xb47, 0xb48}, {0xb4b, 0xb4d},
  116. {0xb56, 0xb57}, {0xb82, 0xb83}, {0xbbe, 0xbc2}, {0xbc6, 0xbc8},
  117. {0xbca, 0xbcd}, {0xbd7, 0xbd7}, {0xc01, 0xc03}, {0xc3e, 0xc44},
  118. {0xc46, 0xc48}, {0xc4a, 0xc4d}, {0xc55, 0xc56}, {0xc82, 0xc83},
  119. {0xcbe, 0xcc4}, {0xcc6, 0xcc8}, {0xcca, 0xccd}, {0xcd5, 0xcd6},
  120. {0xd02, 0xd03}, {0xd3e, 0xd43}, {0xd46, 0xd48}, {0xd4a, 0xd4d},
  121. {0xd57, 0xd57}, {0xe31, 0xe31}, {0xe34, 0xe3a}, {0xe47, 0xe4e},
  122. {0xeb1, 0xeb1}, {0xeb4, 0xeb9}, {0xebb, 0xebc}, {0xec8, 0xecd},
  123. {0xf18, 0xf19}, {0xf35, 0xf35}, {0xf37, 0xf37}, {0xf39, 0xf39},
  124. {0xf3e, 0xf3e}, {0xf3f, 0xf3f}, {0xf71, 0xf84}, {0xf86, 0xf8b},
  125. {0xf90, 0xf95}, {0xf97, 0xf97}, {0xf99, 0xfad}, {0xfb1, 0xfb7},
  126. {0xfb9, 0xfb9}, {0x20d0, 0x20dc}, {0x20e1, 0x20e1}, {0x302a, 0x302f},
  127. {0x3099, 0x3099}, {0x309a, 0x309a}};
  128. const xmlChRangeGroup xmlIsCombiningGroup =
  129. {95, 0, xmlIsCombining_srng, (xmlChLRangePtr)0};
  130. static const xmlChSRange xmlIsDigit_srng[] = { {0x660, 0x669},
  131. {0x6f0, 0x6f9}, {0x966, 0x96f}, {0x9e6, 0x9ef}, {0xa66, 0xa6f},
  132. {0xae6, 0xaef}, {0xb66, 0xb6f}, {0xbe7, 0xbef}, {0xc66, 0xc6f},
  133. {0xce6, 0xcef}, {0xd66, 0xd6f}, {0xe50, 0xe59}, {0xed0, 0xed9},
  134. {0xf20, 0xf29}};
  135. const xmlChRangeGroup xmlIsDigitGroup =
  136. {14, 0, xmlIsDigit_srng, (xmlChLRangePtr)0};
  137. static const xmlChSRange xmlIsExtender_srng[] = { {0x2d0, 0x2d0},
  138. {0x2d1, 0x2d1}, {0x387, 0x387}, {0x640, 0x640}, {0xe46, 0xe46},
  139. {0xec6, 0xec6}, {0x3005, 0x3005}, {0x3031, 0x3035}, {0x309d, 0x309e},
  140. {0x30fc, 0x30fe}};
  141. const xmlChRangeGroup xmlIsExtenderGroup =
  142. {10, 0, xmlIsExtender_srng, (xmlChLRangePtr)0};
  143. static const xmlChSRange xmlIsIdeographic_srng[] = { {0x3007, 0x3007},
  144. {0x3021, 0x3029}, {0x4e00, 0x9fa5}};
  145. const xmlChRangeGroup xmlIsIdeographicGroup =
  146. {3, 0, xmlIsIdeographic_srng, (xmlChLRangePtr)0};
  147. /**
  148. * xmlCharInRange:
  149. * @val: character to be validated
  150. * @rptr: pointer to range to be used to validate
  151. *
  152. * Does a binary search of the range table to determine if char
  153. * is valid
  154. *
  155. * Returns: true if character valid, false otherwise
  156. */
  157. int
  158. xmlCharInRange (unsigned int val, const xmlChRangeGroup *rptr) {
  159. int low, high, mid;
  160. const xmlChSRange *sptr;
  161. const xmlChLRange *lptr;
  162. if (rptr == NULL) return(0);
  163. if (val < 0x10000) { /* is val in 'short' or 'long' array? */
  164. if (rptr->nbShortRange == 0)
  165. return 0;
  166. low = 0;
  167. high = rptr->nbShortRange - 1;
  168. sptr = rptr->shortRange;
  169. while (low <= high) {
  170. mid = (low + high) / 2;
  171. if ((unsigned short) val < sptr[mid].low) {
  172. high = mid - 1;
  173. } else {
  174. if ((unsigned short) val > sptr[mid].high) {
  175. low = mid + 1;
  176. } else {
  177. return 1;
  178. }
  179. }
  180. }
  181. } else {
  182. if (rptr->nbLongRange == 0) {
  183. return 0;
  184. }
  185. low = 0;
  186. high = rptr->nbLongRange - 1;
  187. lptr = rptr->longRange;
  188. while (low <= high) {
  189. mid = (low + high) / 2;
  190. if (val < lptr[mid].low) {
  191. high = mid - 1;
  192. } else {
  193. if (val > lptr[mid].high) {
  194. low = mid + 1;
  195. } else {
  196. return 1;
  197. }
  198. }
  199. }
  200. }
  201. return 0;
  202. }
  203. /**
  204. * xmlIsBaseChar:
  205. * @ch: character to validate
  206. *
  207. * This function is DEPRECATED.
  208. * Use xmlIsBaseChar_ch or xmlIsBaseCharQ instead
  209. *
  210. * Returns true if argument valid, false otherwise
  211. */
  212. int
  213. xmlIsBaseChar(unsigned int ch) {
  214. return(xmlIsBaseCharQ(ch));
  215. }
  216. /**
  217. * xmlIsBlank:
  218. * @ch: character to validate
  219. *
  220. * This function is DEPRECATED.
  221. * Use xmlIsBlank_ch or xmlIsBlankQ instead
  222. *
  223. * Returns true if argument valid, false otherwise
  224. */
  225. int
  226. xmlIsBlank(unsigned int ch) {
  227. return(xmlIsBlankQ(ch));
  228. }
  229. /**
  230. * xmlIsChar:
  231. * @ch: character to validate
  232. *
  233. * This function is DEPRECATED.
  234. * Use xmlIsChar_ch or xmlIsCharQ instead
  235. *
  236. * Returns true if argument valid, false otherwise
  237. */
  238. int
  239. xmlIsChar(unsigned int ch) {
  240. return(xmlIsCharQ(ch));
  241. }
  242. /**
  243. * xmlIsCombining:
  244. * @ch: character to validate
  245. *
  246. * This function is DEPRECATED.
  247. * Use xmlIsCombiningQ instead
  248. *
  249. * Returns true if argument valid, false otherwise
  250. */
  251. int
  252. xmlIsCombining(unsigned int ch) {
  253. return(xmlIsCombiningQ(ch));
  254. }
  255. /**
  256. * xmlIsDigit:
  257. * @ch: character to validate
  258. *
  259. * This function is DEPRECATED.
  260. * Use xmlIsDigit_ch or xmlIsDigitQ instead
  261. *
  262. * Returns true if argument valid, false otherwise
  263. */
  264. int
  265. xmlIsDigit(unsigned int ch) {
  266. return(xmlIsDigitQ(ch));
  267. }
  268. /**
  269. * xmlIsExtender:
  270. * @ch: character to validate
  271. *
  272. * This function is DEPRECATED.
  273. * Use xmlIsExtender_ch or xmlIsExtenderQ instead
  274. *
  275. * Returns true if argument valid, false otherwise
  276. */
  277. int
  278. xmlIsExtender(unsigned int ch) {
  279. return(xmlIsExtenderQ(ch));
  280. }
  281. /**
  282. * xmlIsIdeographic:
  283. * @ch: character to validate
  284. *
  285. * This function is DEPRECATED.
  286. * Use xmlIsIdeographicQ instead
  287. *
  288. * Returns true if argument valid, false otherwise
  289. */
  290. int
  291. xmlIsIdeographic(unsigned int ch) {
  292. return(xmlIsIdeographicQ(ch));
  293. }
  294. /**
  295. * xmlIsPubidChar:
  296. * @ch: character to validate
  297. *
  298. * This function is DEPRECATED.
  299. * Use xmlIsPubidChar_ch or xmlIsPubidCharQ instead
  300. *
  301. * Returns true if argument valid, false otherwise
  302. */
  303. int
  304. xmlIsPubidChar(unsigned int ch) {
  305. return(xmlIsPubidCharQ(ch));
  306. }