jars/icu4j-4_4_2-src/main/tests/collate/src/com/ibm/icu/dev/test/collator/UCAConformanceTest.java

   1 /********************************************************************\r
   2  * COPYRIGHT: \r
   3  * Copyright (c) 2002-2009, International Business Machines Corporation and\r
   4  * others. All Rights Reserved.\r
   5  ********************************************************************/\r
   6 \r
   7 /**\r
   8  * UCAConformanceTest performs conformance tests defined in the data\r
   9  * files. ICU ships with stub data files, as the whole test are too \r
  10  * long. To do the whole test, download the test files.\r
  11  */\r
  12 \r
  13 package com.ibm.icu.dev.test.collator;\r
  14 \r
  15 import java.io.BufferedReader;\r
  16 import java.util.Locale;\r
  17 \r
  18 import com.ibm.icu.dev.test.TestFmwk;\r
  19 import com.ibm.icu.dev.test.TestUtil;\r
  20 import com.ibm.icu.text.CollationKey;\r
  21 import com.ibm.icu.text.Collator;\r
  22 import com.ibm.icu.text.RuleBasedCollator;\r
  23 import com.ibm.icu.text.UTF16;\r
  24 \r
  25 public class UCAConformanceTest extends TestFmwk {\r
  26     \r
  27     /**\r
  28      * @param args\r
  29      */\r
  30     public static void main(String[] args) {\r
  31         new UCAConformanceTest().run(args);\r
  32     }\r
  33     \r
  34     public UCAConformanceTest() {\r
  35     }\r
  36     protected void init()throws Exception{\r
  37         UCA = (RuleBasedCollator)Collator.getInstance(\r
  38                 new Locale("root", "", ""));\r
  39 \r
  40         comparer = new UTF16.StringComparator(true, false, UTF16.StringComparator.FOLD_CASE_DEFAULT);\r
  41     }\r
  42     RuleBasedCollator UCA;\r
  43     RuleBasedCollator rbUCA;\r
  44     UTF16.StringComparator comparer;\r
  45     \r
  46     public void TestTableNonIgnorable() {\r
  47         setCollNonIgnorable(UCA);\r
  48         openTestFile("NON_IGNORABLE");\r
  49         conformanceTest(UCA);\r
  50     }\r
  51     \r
  52     public void TestTableShifted() {\r
  53         setCollShifted(UCA);\r
  54         openTestFile("SHIFTED");\r
  55         conformanceTest(UCA);\r
  56     }\r
  57     \r
  58     public void TestRulesNonIgnorable() {\r
  59         initRbUCA();\r
  60         \r
  61         setCollNonIgnorable(rbUCA);\r
  62         openTestFile("NON_IGNORABLE");\r
  63         conformanceTest(rbUCA);\r
  64     }\r
  65     \r
  66     public void TestRulesShifted() {\r
  67         logln("This test is currently disabled, as it is impossible to "+\r
  68         "wholly represent fractional UCA using tailoring rules.");\r
  69         return;\r
  70         /*        initRbUCA();\r
  71          \r
  72          if(U_SUCCESS(status)) {\r
  73          setCollShifted(rbUCA);\r
  74          openTestFile("SHIFTED");\r
  75          testConformance(rbUCA);\r
  76          }\r
  77          */\r
  78     }\r
  79     BufferedReader in;\r
  80     private void openTestFile(String type)\r
  81     {\r
  82         String collationTest = "CollationTest_";\r
  83         String ext = ".txt";\r
  84         try {\r
  85             if(in != null) {\r
  86                 in.close();\r
  87             }\r
  88         } catch (Exception e) {\r
  89             errln("Could not close the opened file!");\r
  90             return;\r
  91         }\r
  92         try {\r
  93             in = TestUtil.getDataReader(collationTest+type+ext);\r
  94         } catch (Exception e) {\r
  95             try {\r
  96                 in = TestUtil.getDataReader(collationTest+type+"_SHORT"+ext);\r
  97             } catch (Exception e1) {\r
  98                 try {\r
  99                     in = TestUtil.getDataReader(collationTest+type+"_STUB"+ext);\r
 100                     logln( "INFO: Working with the stub file.\n"+\r
 101                             "If you need the full conformance test, please\n"+\r
 102                             "download the appropriate data files from:\n"+\r
 103                     "http://source.icu-project.org/repos/icu/tools/trunk/unicodetools/com/ibm/text/data/");\r
 104                 } catch (Exception e11) {\r
 105                     errln("ERROR: Could not find any of the test files");\r
 106                 }\r
 107             }\r
 108         }\r
 109     }          \r
 110     \r
 111     private void setCollNonIgnorable(RuleBasedCollator coll) \r
 112     {\r
 113         if(coll != null) {\r
 114             coll.setDecomposition(RuleBasedCollator.CANONICAL_DECOMPOSITION);\r
 115             coll.setLowerCaseFirst(false);\r
 116             coll.setCaseLevel(false);\r
 117             coll.setStrength(RuleBasedCollator.TERTIARY);\r
 118             coll.setAlternateHandlingShifted(false);\r
 119         }\r
 120     }\r
 121     \r
 122     private void setCollShifted(RuleBasedCollator coll) \r
 123     {\r
 124         if(coll != null) {\r
 125             coll.setDecomposition(RuleBasedCollator.CANONICAL_DECOMPOSITION);\r
 126             coll.setLowerCaseFirst(false);\r
 127             coll.setCaseLevel(false);\r
 128             coll.setStrength(RuleBasedCollator.QUATERNARY);\r
 129             coll.setAlternateHandlingShifted(true);\r
 130         }\r
 131     }\r
 132     \r
 133     \r
 134     \r
 135     private void initRbUCA() \r
 136     {\r
 137         /*      if(!rbUCA) {\r
 138          UParseError parseError;\r
 139          UChar      *ucarules = buffer;\r
 140          int32_t size = ucol_getRulesEx(UCA, UCOL_FULL_RULES, ucarules, \r
 141          BUFFER_SIZE_);\r
 142          if (size > BUFFER_SIZE_) {\r
 143          ucarules = (UChar *)malloc(size * sizeof(UChar));\r
 144          size = ucol_getRulesEx(UCA, UCOL_FULL_RULES, ucarules, size);\r
 145          }\r
 146          rbUCA = ucol_openRules(ucarules, size, UCOL_DEFAULT, UCOL_TERTIARY, \r
 147          &parseError, &status);\r
 148          if (U_FAILURE(status)) {\r
 149          errln("Failure creating UCA rule-based collator: %s", u_errorName(status));\r
 150          return;\r
 151          }\r
 152          }\r
 153          */\r
 154     }\r
 155     \r
 156     private String parseString(String line) {\r
 157         int i = 0, value;\r
 158         StringBuilder result = new StringBuilder(), buffer = new StringBuilder();\r
 159 \r
 160         for(;;) {\r
 161             while(i < line.length() && Character.isWhitespace(line.charAt(i))) {\r
 162                 i++;\r
 163             }\r
 164             while(i < line.length() && Character.isLetterOrDigit(line.charAt(i))) {\r
 165                 buffer.append(line.charAt(i));\r
 166                 i++;\r
 167             }\r
 168             if(buffer.length() == 0) {\r
 169                 // We hit something that was not whitespace/letter/digit.\r
 170                 // Should be ';' or end of string.\r
 171                 return result.toString();\r
 172             }\r
 173             /* read one code point */\r
 174             value = Integer.parseInt(buffer.toString(), 16);\r
 175             buffer.setLength(0);\r
 176             result.appendCodePoint(value);\r
 177         }\r
 178         \r
 179     }\r
 180     private void conformanceTest(RuleBasedCollator coll) {\r
 181         if(in == null || coll == null) {\r
 182             return;\r
 183         }\r
 184         \r
 185         int lineNo = 0;\r
 186         \r
 187         String line = null, oldLine = null, buffer = null, oldB = null;\r
 188         CollationKey oldSk = null, newSk = null;\r
 189         \r
 190         int res = 0, cmpres = 0, cmpres2 = 0;\r
 191         \r
 192         try {\r
 193             while ((line = in.readLine()) != null) {\r
 194                 lineNo++;\r
 195                 if(line.length() < 3 || line.charAt(0) == '#') {\r
 196                     continue;\r
 197                 }\r
 198                 buffer = parseString(line);\r
 199                 \r
 200                 newSk = coll.getCollationKey(buffer);\r
 201                 if(oldSk != null) {\r
 202                     res = oldSk.compareTo(newSk);\r
 203                     cmpres = coll.compare(oldB, buffer);\r
 204                     cmpres2 = coll.compare(buffer, oldB);\r
 205                     \r
 206                     if(cmpres != -cmpres2) {\r
 207                         errln("Compare result not symmetrical on line "+lineNo);\r
 208                     }\r
 209                     if(((res&0x80000000) != (cmpres&0x80000000)) || (res == 0 && cmpres != 0) || (res != 0 && cmpres == 0)) {\r
 210                         errln("Difference between ucol_strcoll and sortkey compare on line " + lineNo);\r
 211                         logln(oldLine);\r
 212                         logln(line);\r
 213                     }\r
 214                     \r
 215                     if(res > 0) {\r
 216                         errln("Line " + lineNo + " is not greater or equal than previous line");\r
 217                         logln(oldLine);\r
 218                         logln(line);\r
 219                         cmpres = coll.compare(oldB, buffer);\r
 220                     } else if(res == 0) {  // equal \r
 221                         res = comparer.compare(oldB, buffer);\r
 222                         if (res == 0) {\r
 223                             errln("Probable error in test file on line " + lineNo +" (comparing identical strings)");\r
 224                             logln(oldLine);\r
 225                             logln(line);\r
 226                         } else if (res > 0) {\r
 227                             errln("Sortkeys are identical, but code point comapare gives >0 on line " + lineNo);\r
 228                             logln(oldLine);\r
 229                             logln(line);\r
 230                         }\r
 231                     }\r
 232                 }\r
 233                 \r
 234                 oldSk = newSk;\r
 235                 oldB = buffer;\r
 236                 oldLine = line;\r
 237             }\r
 238         } catch (Exception e) {\r
 239             errln("Unexpected exception "+e);\r
 240         }\r
 241     }\r
 242     \r
 243 }\r