/library/java/util/regex/ASCII.java

https://bitbucket.org/chancey/z · Java · 274 lines · 205 code · 40 blank · 29 comment · 3 complexity · 12717b4deeb41a61f8bef59420d7620e MD5 · raw file 

    

  1. /*
    2. 

  2.  * Copyright (c) 1999, 2000, Oracle and/or its affiliates. All rights reserved.
    3. 

  3.  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
    4. 

  4.  *
    5. 

  5.  * This code is free software; you can redistribute it and/or modify it
    6. 

  6.  * under the terms of the GNU General Public License version 2 only, as
    7. 

  7.  * published by the Free Software Foundation.  Oracle designates this
    8. 

  8.  * particular file as subject to the "Classpath" exception as provided
    9. 

  9.  * by Oracle in the LICENSE file that accompanied this code.
    10. 

  10.  *
    11. 

  11.  * This code is distributed in the hope that it will be useful, but WITHOUT
    12. 

  12.  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
    13. 

  13.  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
    14. 

  14.  * version 2 for more details (a copy is included in the LICENSE file that
    15. 

  15.  * accompanied this code).
    16. 

  16.  *
    17. 

  17.  * You should have received a copy of the GNU General Public License version
    18. 

  18.  * 2 along with this work; if not, write to the Free Software Foundation,
    19. 

  19.  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
    20. 

  20.  *
    21. 

  21.  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
    22. 

  22.  * or visit www.oracle.com if you need additional information or have any
    23. 

  23.  * questions.
    24. 

  24.  */
    25. 

  25. 

  26. package java.util.regex;
    27. 

  27. 

  28. 

  29. /**
    30. 

  30.  * Utility class that implements the standard C ctype functionality.
    31. 

  31.  *
    32. 

  32.  * @author Hong Zhang
    33. 

  33.  */
    34. 

  34. 

  35. final class ASCII {
    36. 

  36. 

  37.     static final int UPPER   = 0x00000100;
    38. 

  38. 

  39.     static final int LOWER   = 0x00000200;
    40. 

  40. 

  41.     static final int DIGIT   = 0x00000400;
    42. 

  42. 

  43.     static final int SPACE   = 0x00000800;
    44. 

  44. 

  45.     static final int PUNCT   = 0x00001000;
    46. 

  46. 

  47.     static final int CNTRL   = 0x00002000;
    48. 

  48. 

  49.     static final int BLANK   = 0x00004000;
    50. 

  50. 

  51.     static final int HEX     = 0x00008000;
    52. 

  52. 

  53.     static final int UNDER   = 0x00010000;
    54. 

  54. 

  55.     static final int ASCII   = 0x0000FF00;
    56. 

  56. 

  57.     static final int ALPHA   = (UPPER|LOWER);
    58. 

  58. 

  59.     static final int ALNUM   = (UPPER|LOWER|DIGIT);
    60. 

  60. 

  61.     static final int GRAPH   = (PUNCT|UPPER|LOWER|DIGIT);
    62. 

  62. 

  63.     static final int WORD    = (UPPER|LOWER|UNDER|DIGIT);
    64. 

  64. 

  65.     static final int XDIGIT  = (HEX);
    66. 

  66. 

  67.     private static final int[] ctype = new int[] {
    68. 

  68.         CNTRL,                  /* 00 (NUL) */
    69. 

  69.         CNTRL,                  /* 01 (SOH) */
    70. 

  70.         CNTRL,                  /* 02 (STX) */
    71. 

  71.         CNTRL,                  /* 03 (ETX) */
    72. 

  72.         CNTRL,                  /* 04 (EOT) */
    73. 

  73.         CNTRL,                  /* 05 (ENQ) */
    74. 

  74.         CNTRL,                  /* 06 (ACK) */
    75. 

  75.         CNTRL,                  /* 07 (BEL) */
    76. 

  76.         CNTRL,                  /* 08 (BS)  */
    77. 

  77.         SPACE+CNTRL+BLANK,      /* 09 (HT)  */
    78. 

  78.         SPACE+CNTRL,            /* 0A (LF)  */
    79. 

  79.         SPACE+CNTRL,            /* 0B (VT)  */
    80. 

  80.         SPACE+CNTRL,            /* 0C (FF)  */
    81. 

  81.         SPACE+CNTRL,            /* 0D (CR)  */
    82. 

  82.         CNTRL,                  /* 0E (SI)  */
    83. 

  83.         CNTRL,                  /* 0F (SO)  */
    84. 

  84.         CNTRL,                  /* 10 (DLE) */
    85. 

  85.         CNTRL,                  /* 11 (DC1) */
    86. 

  86.         CNTRL,                  /* 12 (DC2) */
    87. 

  87.         CNTRL,                  /* 13 (DC3) */
    88. 

  88.         CNTRL,                  /* 14 (DC4) */
    89. 

  89.         CNTRL,                  /* 15 (NAK) */
    90. 

  90.         CNTRL,                  /* 16 (SYN) */
    91. 

  91.         CNTRL,                  /* 17 (ETB) */
    92. 

  92.         CNTRL,                  /* 18 (CAN) */
    93. 

  93.         CNTRL,                  /* 19 (EM)  */
    94. 

  94.         CNTRL,                  /* 1A (SUB) */
    95. 

  95.         CNTRL,                  /* 1B (ESC) */
    96. 

  96.         CNTRL,                  /* 1C (FS)  */
    97. 

  97.         CNTRL,                  /* 1D (GS)  */
    98. 

  98.         CNTRL,                  /* 1E (RS)  */
    99. 

  99.         CNTRL,                  /* 1F (US)  */
    100. 

  100.         SPACE+BLANK,            /* 20 SPACE */
    101. 

  101.         PUNCT,                  /* 21 !     */
    102. 

  102.         PUNCT,                  /* 22 "     */
    103. 

  103.         PUNCT,                  /* 23 #     */
    104. 

  104.         PUNCT,                  /* 24 $     */
    105. 

  105.         PUNCT,                  /* 25 %     */
    106. 

  106.         PUNCT,                  /* 26 &     */
    107. 

  107.         PUNCT,                  /* 27 '     */
    108. 

  108.         PUNCT,                  /* 28 (     */
    109. 

  109.         PUNCT,                  /* 29 )     */
    110. 

  110.         PUNCT,                  /* 2A *     */
    111. 

  111.         PUNCT,                  /* 2B +     */
    112. 

  112.         PUNCT,                  /* 2C ,     */
    113. 

  113.         PUNCT,                  /* 2D -     */
    114. 

  114.         PUNCT,                  /* 2E .     */
    115. 

  115.         PUNCT,                  /* 2F /     */
    116. 

  116.         DIGIT+HEX+0,            /* 30 0     */
    117. 

  117.         DIGIT+HEX+1,            /* 31 1     */
    118. 

  118.         DIGIT+HEX+2,            /* 32 2     */
    119. 

  119.         DIGIT+HEX+3,            /* 33 3     */
    120. 

  120.         DIGIT+HEX+4,            /* 34 4     */
    121. 

  121.         DIGIT+HEX+5,            /* 35 5     */
    122. 

  122.         DIGIT+HEX+6,            /* 36 6     */
    123. 

  123.         DIGIT+HEX+7,            /* 37 7     */
    124. 

  124.         DIGIT+HEX+8,            /* 38 8     */
    125. 

  125.         DIGIT+HEX+9,            /* 39 9     */
    126. 

  126.         PUNCT,                  /* 3A :     */
    127. 

  127.         PUNCT,                  /* 3B ;     */
    128. 

  128.         PUNCT,                  /* 3C <     */
    129. 

  129.         PUNCT,                  /* 3D =     */
    130. 

  130.         PUNCT,                  /* 3E >     */
    131. 

  131.         PUNCT,                  /* 3F ?     */
    132. 

  132.         PUNCT,                  /* 40 @     */
    133. 

  133.         UPPER+HEX+10,           /* 41 A     */
    134. 

  134.         UPPER+HEX+11,           /* 42 B     */
    135. 

  135.         UPPER+HEX+12,           /* 43 C     */
    136. 

  136.         UPPER+HEX+13,           /* 44 D     */
    137. 

  137.         UPPER+HEX+14,           /* 45 E     */
    138. 

  138.         UPPER+HEX+15,           /* 46 F     */
    139. 

  139.         UPPER+16,               /* 47 G     */
    140. 

  140.         UPPER+17,               /* 48 H     */
    141. 

  141.         UPPER+18,               /* 49 I     */
    142. 

  142.         UPPER+19,               /* 4A J     */
    143. 

  143.         UPPER+20,               /* 4B K     */
    144. 

  144.         UPPER+21,               /* 4C L     */
    145. 

  145.         UPPER+22,               /* 4D M     */
    146. 

  146.         UPPER+23,               /* 4E N     */
    147. 

  147.         UPPER+24,               /* 4F O     */
    148. 

  148.         UPPER+25,               /* 50 P     */
    149. 

  149.         UPPER+26,               /* 51 Q     */
    150. 

  150.         UPPER+27,               /* 52 R     */
    151. 

  151.         UPPER+28,               /* 53 S     */
    152. 

  152.         UPPER+29,               /* 54 T     */
    153. 

  153.         UPPER+30,               /* 55 U     */
    154. 

  154.         UPPER+31,               /* 56 V     */
    155. 

  155.         UPPER+32,               /* 57 W     */
    156. 

  156.         UPPER+33,               /* 58 X     */
    157. 

  157.         UPPER+34,               /* 59 Y     */
    158. 

  158.         UPPER+35,               /* 5A Z     */
    159. 

  159.         PUNCT,                  /* 5B [     */
    160. 

  160.         PUNCT,                  /* 5C \     */
    161. 

  161.         PUNCT,                  /* 5D ]     */
    162. 

  162.         PUNCT,                  /* 5E ^     */
    163. 

  163.         PUNCT|UNDER,            /* 5F _     */
    164. 

  164.         PUNCT,                  /* 60 `     */
    165. 

  165.         LOWER+HEX+10,           /* 61 a     */
    166. 

  166.         LOWER+HEX+11,           /* 62 b     */
    167. 

  167.         LOWER+HEX+12,           /* 63 c     */
    168. 

  168.         LOWER+HEX+13,           /* 64 d     */
    169. 

  169.         LOWER+HEX+14,           /* 65 e     */
    170. 

  170.         LOWER+HEX+15,           /* 66 f     */
    171. 

  171.         LOWER+16,               /* 67 g     */
    172. 

  172.         LOWER+17,               /* 68 h     */
    173. 

  173.         LOWER+18,               /* 69 i     */
    174. 

  174.         LOWER+19,               /* 6A j     */
    175. 

  175.         LOWER+20,               /* 6B k     */
    176. 

  176.         LOWER+21,               /* 6C l     */
    177. 

  177.         LOWER+22,               /* 6D m     */
    178. 

  178.         LOWER+23,               /* 6E n     */
    179. 

  179.         LOWER+24,               /* 6F o     */
    180. 

  180.         LOWER+25,               /* 70 p     */
    181. 

  181.         LOWER+26,               /* 71 q     */
    182. 

  182.         LOWER+27,               /* 72 r     */
    183. 

  183.         LOWER+28,               /* 73 s     */
    184. 

  184.         LOWER+29,               /* 74 t     */
    185. 

  185.         LOWER+30,               /* 75 u     */
    186. 

  186.         LOWER+31,               /* 76 v     */
    187. 

  187.         LOWER+32,               /* 77 w     */
    188. 

  188.         LOWER+33,               /* 78 x     */
    189. 

  189.         LOWER+34,               /* 79 y     */
    190. 

  190.         LOWER+35,               /* 7A z     */
    191. 

  191.         PUNCT,                  /* 7B {     */
    192. 

  192.         PUNCT,                  /* 7C |     */
    193. 

  193.         PUNCT,                  /* 7D }     */
    194. 

  194.         PUNCT,                  /* 7E ~     */
    195. 

  195.         CNTRL,                  /* 7F (DEL) */
    196. 

  196.     };
    197. 

  197. 

  198.     static int getType(int ch) {
    199. 

  199.         return ((ch & 0xFFFFFF80) == 0 ? ctype[ch] : 0);
    200. 

  200.     }
    201. 

  201. 

  202.     static boolean isType(int ch, int type) {
    203. 

  203.         return (getType(ch) & type) != 0;
    204. 

  204.     }
    205. 

  205. 

  206.     static boolean isAscii(int ch) {
    207. 

  207.         return ((ch & 0xFFFFFF80) == 0);
    208. 

  208.     }
    209. 

  209. 

  210.     static boolean isAlpha(int ch) {
    211. 

  211.         return isType(ch, ALPHA);
    212. 

  212.     }
    213. 

  213. 

  214.     static boolean isDigit(int ch) {
    215. 

  215.         return ((ch-'0')|('9'-ch)) >= 0;
    216. 

  216.     }
    217. 

  217. 

  218.     static boolean isAlnum(int ch) {
    219. 

  219.         return isType(ch, ALNUM);
    220. 

  220.     }
    221. 

  221. 

  222.     static boolean isGraph(int ch) {
    223. 

  223.         return isType(ch, GRAPH);
    224. 

  224.     }
    225. 

  225. 

  226.     static boolean isPrint(int ch) {
    227. 

  227.         return ((ch-0x20)|(0x7E-ch)) >= 0;
    228. 

  228.     }
    229. 

  229. 

  230.     static boolean isPunct(int ch) {
    231. 

  231.         return isType(ch, PUNCT);
    232. 

  232.     }
    233. 

  233. 

  234.     static boolean isSpace(int ch) {
    235. 

  235.         return isType(ch, SPACE);
    236. 

  236.     }
    237. 

  237. 

  238.     static boolean isHexDigit(int ch) {
    239. 

  239.         return isType(ch, HEX);
    240. 

  240.     }
    241. 

  241. 

  242.     static boolean isOctDigit(int ch) {
    243. 

  243.         return ((ch-'0')|('7'-ch)) >= 0;
    244. 

  244.     }
    245. 

  245. 

  246.     static boolean isCntrl(int ch) {
    247. 

  247.         return isType(ch, CNTRL);
    248. 

  248.     }
    249. 

  249. 

  250.     static boolean isLower(int ch) {
    251. 

  251.         return ((ch-'a')|('z'-ch)) >= 0;
    252. 

  252.     }
    253. 

  253. 

  254.     static boolean isUpper(int ch) {
    255. 

  255.         return ((ch-'A')|('Z'-ch)) >= 0;
    256. 

  256.     }
    257. 

  257. 

  258.     static boolean isWord(int ch) {
    259. 

  259.         return isType(ch, WORD);
    260. 

  260.     }
    261. 

  261. 

  262.     static int toDigit(int ch) {
    263. 

  263.         return (ctype[ch & 0x7F] & 0x3F);
    264. 

  264.     }
    265. 

  265. 

  266.     static int toLower(int ch) {
    267. 

  267.         return isUpper(ch) ? (ch + 0x20) : ch;
    268. 

  268.     }
    269. 

  269. 

  270.     static int toUpper(int ch) {
    271. 

  271.         return isLower(ch) ? (ch - 0x20) : ch;
    272. 

  272.     }
    273. 

  273. 

  274. }
    275.