/interpreter/tags/at2-build190607/src/edu/vub/util/regexp/RETokenNamedProperty.java

http://ambienttalk.googlecode.com/ · Java · 311 lines · 210 code · 35 blank · 66 comment · 49 complexity · af77e9b84d789bd8c606cdcb5bce3f96 MD5 · raw file

  1. /* gnu/regexp/RETokenNamedProperty.java
  2. Copyright (C) 2006 Free Software Foundation, Inc.
  3. This file is part of GNU Classpath.
  4. GNU Classpath is free software; you can redistribute it and/or modify
  5. it under the terms of the GNU General Public License as published by
  6. the Free Software Foundation; either version 2, or (at your option)
  7. any later version.
  8. GNU Classpath is distributed in the hope that it will be useful, but
  9. WITHOUT ANY WARRANTY; without even the implied warranty of
  10. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  11. General Public License for more details.
  12. You should have received a copy of the GNU General Public License
  13. along with GNU Classpath; see the file COPYING. If not, write to the
  14. Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  15. 02110-1301 USA.
  16. Linking this library statically or dynamically with other modules is
  17. making a combined work based on this library. Thus, the terms and
  18. conditions of the GNU General Public License cover the whole
  19. combination.
  20. As a special exception, the copyright holders of this library give you
  21. permission to link this library with independent modules to produce an
  22. executable, regardless of the license terms of these independent
  23. modules, and to copy and distribute the resulting executable under
  24. terms of your choice, provided that you also meet, for each linked
  25. independent module, the terms and conditions of the license of that
  26. module. An independent module is a module which is not derived from
  27. or based on this library. If you modify this library, you may extend
  28. this exception to your version of the library, but you are not
  29. obligated to do so. If you do not wish to do so, delete this
  30. exception statement from your version. */
  31. package edu.vub.util.regexp;
  32. final class RETokenNamedProperty extends REToken {
  33. String name;
  34. boolean insens;
  35. boolean negate;
  36. Handler handler;
  37. // Grouped properties
  38. static final byte[] LETTER = new byte[]
  39. { Character.LOWERCASE_LETTER,
  40. Character.UPPERCASE_LETTER,
  41. Character.TITLECASE_LETTER,
  42. Character.MODIFIER_LETTER,
  43. Character.OTHER_LETTER };
  44. static final byte[] MARK = new byte[]
  45. { Character.NON_SPACING_MARK,
  46. Character.COMBINING_SPACING_MARK,
  47. Character.ENCLOSING_MARK };
  48. static final byte[] SEPARATOR = new byte[]
  49. { Character.SPACE_SEPARATOR,
  50. Character.LINE_SEPARATOR,
  51. Character.PARAGRAPH_SEPARATOR };
  52. static final byte[] SYMBOL = new byte[]
  53. { Character.MATH_SYMBOL,
  54. Character.CURRENCY_SYMBOL,
  55. Character.MODIFIER_SYMBOL,
  56. Character.OTHER_SYMBOL };
  57. static final byte[] NUMBER = new byte[]
  58. { Character.DECIMAL_DIGIT_NUMBER,
  59. Character.LETTER_NUMBER,
  60. Character.OTHER_NUMBER };
  61. static final byte[] PUNCTUATION = new byte[]
  62. { Character.DASH_PUNCTUATION,
  63. Character.START_PUNCTUATION,
  64. Character.END_PUNCTUATION,
  65. Character.CONNECTOR_PUNCTUATION,
  66. Character.OTHER_PUNCTUATION }; // ,
  67. // Character.INITIAL_QUOTE_PUNCTUATION,
  68. // Character.FINAL_QUOTE_PUNCTUATION};
  69. static final byte[] OTHER = new byte[]
  70. { Character.CONTROL,
  71. Character.FORMAT,
  72. Character.PRIVATE_USE,
  73. Character.SURROGATE,
  74. Character.UNASSIGNED };
  75. RETokenNamedProperty(int subIndex, String name, boolean insens, boolean negate) throws REException {
  76. super(subIndex);
  77. this.name = name;
  78. this.insens = insens;
  79. this.negate = negate;
  80. handler = getHandler(name);
  81. }
  82. int getMinimumLength() {
  83. return 1;
  84. }
  85. int getMaximumLength() {
  86. return 1;
  87. }
  88. boolean match(CharIndexed input, REMatch mymatch) {
  89. char ch = input.charAt(mymatch.index);
  90. if (ch == CharIndexed.OUT_OF_BOUNDS)
  91. return false;
  92. boolean retval = handler.includes(ch);
  93. if (insens) {
  94. retval = retval ||
  95. handler.includes(Character.toUpperCase(ch)) ||
  96. handler.includes(Character.toLowerCase(ch));
  97. }
  98. if (negate) retval = !retval;
  99. if (retval) {
  100. ++mymatch.index;
  101. return next(input, mymatch);
  102. }
  103. else return false;
  104. }
  105. void dump(StringBuffer os) {
  106. os.append("\\")
  107. .append(negate ? "P" : "p")
  108. .append("{" + name + "}");
  109. }
  110. private abstract static class Handler {
  111. abstract boolean includes(char c);
  112. }
  113. private Handler getHandler(String name) throws REException {
  114. if (name.equals("Lower") ||
  115. name.equals("Upper") ||
  116. // name.equals("ASCII") ||
  117. name.equals("Alpha") ||
  118. name.equals("Digit") ||
  119. name.equals("Alnum") ||
  120. name.equals("Punct") ||
  121. name.equals("Graph") ||
  122. name.equals("Print") ||
  123. name.equals("Blank") ||
  124. name.equals("Cntrl") ||
  125. name.equals("XDigit") ||
  126. name.equals("Space") ) {
  127. return new POSIXHandler(name);
  128. }
  129. // if (name.startsWith("In")) {
  130. // try {
  131. // name = name.substring(2);
  132. // Character.UnicodeBlock block = Character.UnicodeBlock.forName(name);
  133. // return new UnicodeBlockHandler(block);
  134. // }
  135. // catch (IllegalArgumentException e) {
  136. // throw new REException("Invalid Unicode block name: " + name, REException.REG_ESCAPE, 0);
  137. // }
  138. // }
  139. if (name.startsWith("Is")) {
  140. name = name.substring(2);
  141. }
  142. // "grouped properties"
  143. if (name.equals("L"))
  144. return new UnicodeCategoriesHandler(LETTER);
  145. if (name.equals("M"))
  146. return new UnicodeCategoriesHandler(MARK);
  147. if (name.equals("Z"))
  148. return new UnicodeCategoriesHandler(SEPARATOR);
  149. if (name.equals("S"))
  150. return new UnicodeCategoriesHandler(SYMBOL);
  151. if (name.equals("N"))
  152. return new UnicodeCategoriesHandler(NUMBER);
  153. if (name.equals("P"))
  154. return new UnicodeCategoriesHandler(PUNCTUATION);
  155. if (name.equals("C"))
  156. return new UnicodeCategoriesHandler(OTHER);
  157. if (name.equals("Mc"))
  158. return new UnicodeCategoryHandler(Character.COMBINING_SPACING_MARK);
  159. if (name.equals("Pc"))
  160. return new UnicodeCategoryHandler(Character.CONNECTOR_PUNCTUATION);
  161. if (name.equals("Cc"))
  162. return new UnicodeCategoryHandler(Character.CONTROL);
  163. if (name.equals("Sc"))
  164. return new UnicodeCategoryHandler(Character.CURRENCY_SYMBOL);
  165. if (name.equals("Pd"))
  166. return new UnicodeCategoryHandler(Character.DASH_PUNCTUATION);
  167. if (name.equals("Nd"))
  168. return new UnicodeCategoryHandler(Character.DECIMAL_DIGIT_NUMBER);
  169. if (name.equals("Me"))
  170. return new UnicodeCategoryHandler(Character.ENCLOSING_MARK);
  171. if (name.equals("Pe"))
  172. return new UnicodeCategoryHandler(Character.END_PUNCTUATION);
  173. // if (name.equals("Pf"))
  174. // return new UnicodeCategoryHandler(Character.FINAL_QUOTE_PUNCTUATION);
  175. if (name.equals("Cf"))
  176. return new UnicodeCategoryHandler(Character.FORMAT);
  177. // if (name.equals("Pi"))
  178. // return new UnicodeCategoryHandler(Character.INITIAL_QUOTE_PUNCTUATION);
  179. if (name.equals("Nl"))
  180. return new UnicodeCategoryHandler(Character.LETTER_NUMBER);
  181. if (name.equals("Zl"))
  182. return new UnicodeCategoryHandler(Character.LINE_SEPARATOR);
  183. if (name.equals("Ll"))
  184. return new UnicodeCategoryHandler(Character.LOWERCASE_LETTER);
  185. if (name.equals("Sm"))
  186. return new UnicodeCategoryHandler(Character.MATH_SYMBOL);
  187. if (name.equals("Lm"))
  188. return new UnicodeCategoryHandler(Character.MODIFIER_LETTER);
  189. if (name.equals("Sk"))
  190. return new UnicodeCategoryHandler(Character.MODIFIER_SYMBOL);
  191. if (name.equals("Mn"))
  192. return new UnicodeCategoryHandler(Character.NON_SPACING_MARK);
  193. if (name.equals("Lo"))
  194. return new UnicodeCategoryHandler(Character.OTHER_LETTER);
  195. if (name.equals("No"))
  196. return new UnicodeCategoryHandler(Character.OTHER_NUMBER);
  197. if (name.equals("Po"))
  198. return new UnicodeCategoryHandler(Character.OTHER_PUNCTUATION);
  199. if (name.equals("So"))
  200. return new UnicodeCategoryHandler(Character.OTHER_SYMBOL);
  201. if (name.equals("Zp"))
  202. return new UnicodeCategoryHandler(Character.PARAGRAPH_SEPARATOR);
  203. if (name.equals("Co"))
  204. return new UnicodeCategoryHandler(Character.PRIVATE_USE);
  205. if (name.equals("Zs"))
  206. return new UnicodeCategoryHandler(Character.SPACE_SEPARATOR);
  207. if (name.equals("Ps"))
  208. return new UnicodeCategoryHandler(Character.START_PUNCTUATION);
  209. if (name.equals("Cs"))
  210. return new UnicodeCategoryHandler(Character.SURROGATE);
  211. if (name.equals("Lt"))
  212. return new UnicodeCategoryHandler(Character.TITLECASE_LETTER);
  213. if (name.equals("Cn"))
  214. return new UnicodeCategoryHandler(Character.UNASSIGNED);
  215. if (name.equals("Lu"))
  216. return new UnicodeCategoryHandler(Character.UPPERCASE_LETTER);
  217. throw new REException("unsupported name " + name, REException.REG_ESCAPE, 0);
  218. }
  219. private static class POSIXHandler extends Handler {
  220. private RETokenPOSIX retoken;
  221. private REMatch mymatch = new REMatch(0, 0, 0);
  222. private char[] chars = new char[1];
  223. private CharIndexedCharArray ca = new CharIndexedCharArray(chars, 0);
  224. POSIXHandler(String name) {
  225. int posixId = RETokenPOSIX.intValue(name.toLowerCase());
  226. if (posixId != -1)
  227. retoken = new RETokenPOSIX(0, posixId, false, false);
  228. else
  229. throw new RuntimeException("Unknown posix ID: " + name);
  230. }
  231. boolean includes(char c) {
  232. chars[0] = c;
  233. mymatch.index = 0;
  234. return retoken.match(ca, mymatch);
  235. }
  236. }
  237. private static class UnicodeCategoryHandler extends Handler {
  238. UnicodeCategoryHandler(byte category) {
  239. this.category = (int) category;
  240. }
  241. private int category;
  242. boolean includes(char c) {
  243. return Character.getType(c) == category;
  244. }
  245. }
  246. private static class UnicodeCategoriesHandler extends Handler {
  247. UnicodeCategoriesHandler(byte[] categories) {
  248. this.categories = categories;
  249. }
  250. private byte[] categories;
  251. boolean includes(char c) {
  252. int category = Character.getType(c);
  253. for (int i = 0; i < categories.length; i++)
  254. if (category == categories[i])
  255. return true;
  256. return false;
  257. }
  258. }
  259. // Backport from JDK 1.4 to 1.3
  260. // private static class UnicodeBlockHandler extends Handler {
  261. // public UnicodeBlockHandler(Character.UnicodeBlock block) {
  262. // this.block = block;
  263. // }
  264. // private Character.UnicodeBlock block;
  265. // public boolean includes(char c) {
  266. // Character.UnicodeBlock cblock = Character.UnicodeBlock.of(c);
  267. // return (cblock != null && cblock.equals(block));
  268. // }
  269. // }
  270. }