scuffed-code/icu4c/source/test/cintltst/citertst.c

/********************************************************************
 * COPYRIGHT:
 * Copyright (c) 1997-2001, International Business Machines Corporation and
 * others. All Rights Reserved.
 ********************************************************************/
/********************************************************************************
*
* File CITERTST.C
*
* Modification History:
* Date      Name               Description
*           Madhu Katragadda   Ported for C API
* 02/19/01  synwee             Modified test case for new collation iterator
*********************************************************************************/
/*
 * Collation Iterator tests.
 * (Let me reiterate my position...)
 */

#include "unicode/utypes.h"
#include "unicode/ucol.h"
#include "unicode/uloc.h"
#include "unicode/uchar.h"
#include "unicode/ustring.h"
#include "cmemory.h"
#include "cintltst.h"
#include "citertst.h"
#include "ccolltst.h"
#include "filestrm.h"
#include "cstring.h"
#include "ucol_imp.h"

void addCollIterTest(TestNode** root)
{
    addTest(root, &TestPrevious, "tscoll/citertst/TestPrevious");
    addTest(root, &TestOffset, "tscoll/citertst/TestOffset");
    addTest(root, &TestSetText, "tscoll/citertst/TestSetText");
    addTest(root, &TestMaxExpansion, "tscoll/citertst/TestMaxExpansion");
    addTest(root, &TestUnicodeChar, "tscoll/citertst/TestUnicodeChar");
    addTest(root, &TestNormalizedUnicodeChar,
                                "tscoll/citertst/TestNormalizedUnicodeChar");
    addTest(root, &TestNormalization, "tscoll/citertst/TestNormalization");
    addTest(root, &TestBug672, "tscoll/citertst/TestBug672");
    addTest(root, &TestBug672Normalize, "tscoll/citertst/TestBug672Normalize");
    addTest(root, &TestSmallBuffer, "tscoll/citertst/TestSmallBuffer");
    addTest(root, &TestCEs, "tscoll/citertst/TestCEs");
    addTest(root, &TestDiscontiguos, "tscoll/citertst/TestDiscontiguos");
}

/* The locales we support */

static const char * LOCALES[] = {"en_AU", "en_BE", "en_CA"};

static void TestBug672() {
    UErrorCode  status = U_ZERO_ERROR;
    UChar       pattern[20];
    UChar       text[50];
    int         i;
    int         result[3][3];

    u_uastrcpy(pattern, "resume");
    u_uastrcpy(text, "Time to resume updating my resume.");

    for (i = 0; i < 3; ++ i) {
        UCollator          *coll = ucol_open(LOCALES[i], &status);
        UCollationElements *pitr = ucol_openElements(coll, pattern, -1,
                                                     &status);
        UCollationElements *titer = ucol_openElements(coll, text, -1,
                                                     &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: in creation of either the collator or the collation iterator :%s\n",
                    myErrorName(status));
            return;
        }

        log_verbose("locale tested %s\n", LOCALES[i]);

        while (ucol_next(pitr, &status) != UCOL_NULLORDER &&
               U_SUCCESS(status)) {
        }
        if (U_FAILURE(status)) {
            log_err("ERROR: reversing collation iterator :%s\n",
                    myErrorName(status));
            return;
        }
        ucol_reset(pitr);

        ucol_setOffset(titer, u_strlen(pattern), &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: setting offset in collator :%s\n",
                    myErrorName(status));
            return;
        }
        result[i][0] = ucol_getOffset(titer);
        log_verbose("Text iterator set to offset %d\n", result[i][0]);

        /* Use previous() */
        ucol_previous(titer, &status);
        result[i][1] = ucol_getOffset(titer);
        log_verbose("Current offset %d after previous\n", result[i][1]);

        /* Add one to index */
        log_verbose("Adding one to current offset...\n");
        ucol_setOffset(titer, ucol_getOffset(titer) + 1, &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: setting offset in collator :%s\n",
                    myErrorName(status));
            return;
        }
        result[i][2] = ucol_getOffset(titer);
        log_verbose("Current offset in text = %d\n", result[i][2]);
        ucol_closeElements(pitr);
        ucol_closeElements(titer);
        ucol_close(coll);
    }

    if (uprv_memcmp(result[0], result[1], 3) != 0 ||
        uprv_memcmp(result[1], result[2], 3) != 0) {
        log_err("ERROR: Different locales have different offsets at the same character\n");
    }
}



/*  Running this test with normalization enabled showed up a bug in the incremental
    normalization code. */
static void TestBug672Normalize() {
    UErrorCode  status = U_ZERO_ERROR;
    UChar       pattern[20];
    UChar       text[50];
    int         i;
    int         result[3][3];

    u_uastrcpy(pattern, "resume");
    u_uastrcpy(text, "Time to resume updating my resume.");

    for (i = 0; i < 3; ++ i) {
        UCollator          *coll = ucol_open(LOCALES[i], &status);
        UCollationElements *pitr = NULL;
        UCollationElements *titer = NULL;

        ucol_setAttribute(coll, UCOL_NORMALIZATION_MODE, UCOL_ON, &status);

        pitr = ucol_openElements(coll, pattern, -1, &status);
        titer = ucol_openElements(coll, text, -1, &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: in creation of either the collator or the collation iterator :%s\n",
                    myErrorName(status));
            return;
        }

        log_verbose("locale tested %s\n", LOCALES[i]);

        while (ucol_next(pitr, &status) != UCOL_NULLORDER &&
               U_SUCCESS(status)) {
        }
        if (U_FAILURE(status)) {
            log_err("ERROR: reversing collation iterator :%s\n",
                    myErrorName(status));
            return;
        }
        ucol_reset(pitr);

        ucol_setOffset(titer, u_strlen(pattern), &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: setting offset in collator :%s\n",
                    myErrorName(status));
            return;
        }
        result[i][0] = ucol_getOffset(titer);
        log_verbose("Text iterator set to offset %d\n", result[i][0]);

        /* Use previous() */
        ucol_previous(titer, &status);
        result[i][1] = ucol_getOffset(titer);
        log_verbose("Current offset %d after previous\n", result[i][1]);

        /* Add one to index */
        log_verbose("Adding one to current offset...\n");
        ucol_setOffset(titer, ucol_getOffset(titer) + 1, &status);
        if (U_FAILURE(status)) {
            log_err("ERROR: setting offset in collator :%s\n",
                    myErrorName(status));
            return;
        }
        result[i][2] = ucol_getOffset(titer);
        log_verbose("Current offset in text = %d\n", result[i][2]);
        ucol_closeElements(pitr);
        ucol_closeElements(titer);
        ucol_close(coll);
    }

    if (uprv_memcmp(result[0], result[1], 3) != 0 ||
        uprv_memcmp(result[1], result[2], 3) != 0) {
        log_err("ERROR: Different locales have different offsets at the same character\n");
    }
}




/**
 * Test for CollationElementIterator previous and next for the whole set of
 * unicode characters.
 */
static void TestUnicodeChar()
{
    UChar source[0x100];
    UCollator *en_us;
    UCollationElements *iter;
    UErrorCode status = U_ZERO_ERROR;
    UChar codepoint;

    UChar *test;
    en_us = ucol_open("en_US", &status);
    if (U_FAILURE(status)){
       log_err("ERROR: in creation of collation data using ucol_open()\n %s\n",
              myErrorName(status));
       return;
    }

    for (codepoint = 1; codepoint < 0xFFFE;)
    {
      test = source;

      while (codepoint % 0xFF != 0)
      {
        if (u_isdefined(codepoint))
          *(test ++) = codepoint;
        codepoint ++;
      }

      if (u_isdefined(codepoint))
        *(test ++) = codepoint;

      if (codepoint != 0xFFFF)
        codepoint ++;

      *test = 0;
      iter=ucol_openElements(en_us, source, u_strlen(source), &status);
      if(U_FAILURE(status)){
          log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
              myErrorName(status));
          ucol_close(en_us);
          return;
      }
      /* A basic test to see if it's working at all */
      log_verbose("codepoint testing %x\n", codepoint);
      backAndForth(iter);
      ucol_closeElements(iter);

      /* null termination test */
      iter=ucol_openElements(en_us, source, -1, &status);
      if(U_FAILURE(status)){
          log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
              myErrorName(status));
          ucol_close(en_us);
          return;
      }
      /* A basic test to see if it's working at all */
      backAndForth(iter);
      ucol_closeElements(iter);
    }

    ucol_close(en_us);
}

/**
 * Test for CollationElementIterator previous and next for the whole set of
 * unicode characters with normalization on.
 */
static void TestNormalizedUnicodeChar()
{
    UChar source[0x100];
    UCollator *th_th;
    UCollationElements *iter;
    UErrorCode status = U_ZERO_ERROR;
    UChar codepoint;

    UChar *test;
    /* thai should have normalization on */
    th_th = ucol_open("th_TH", &status);
    if (U_FAILURE(status)){
        log_err("ERROR: in creation of thai collation using ucol_open()\n %s\n",
              myErrorName(status));
        return;
    }

    for (codepoint = 1; codepoint < 0xFFFE;)
    {
      test = source;

      while (codepoint % 0xFF != 0)
      {
        if (u_isdefined(codepoint))
          *(test ++) = codepoint;
        codepoint ++;
      }

      if (u_isdefined(codepoint))
        *(test ++) = codepoint;

      if (codepoint != 0xFFFF)
        codepoint ++;

      *test = 0;
      iter=ucol_openElements(th_th, source, u_strlen(source), &status);
      if(U_FAILURE(status)){
          log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
              myErrorName(status));
            ucol_close(th_th);
          return;
      }
        
      backAndForth(iter);
      ucol_closeElements(iter);

      iter=ucol_openElements(th_th, source, -1, &status);
      if(U_FAILURE(status)){
          log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
              myErrorName(status));
            ucol_close(th_th);
          return;
      }

      backAndForth(iter);
      ucol_closeElements(iter);
    }

    ucol_close(th_th);
}

/**
* Test the incremental normalization
*/
static void TestNormalization()
{
          UErrorCode          status = U_ZERO_ERROR;
    const char               *str    = 
                            "&a < \\u0300\\u0315 < A\\u0300\\u0315 < \\u0316\\u0315B < \\u0316\\u0300\\u0315";
          UCollator          *coll;
          UChar               rule[50];
          int                 rulelen = u_unescape(str, rule, 50);
          int                 count = 0;
    const char                *testdata[] =
                        {"\\u1ED9", "o\\u0323\\u0302",
                        "\\u0300\\u0315", "\\u0315\\u0300",
                        "A\\u0300\\u0315B", "A\\u0315\\u0300B",
                        "A\\u0316\\u0315B", "A\\u0315\\u0316B",
                        "\\u0316\\u0300\\u0315", "\\u0315\\u0300\\u0316",
                        "A\\u0316\\u0300\\u0315B", "A\\u0315\\u0300\\u0316B",
                        "\\u0316\\u0315\\u0300", "A\\u0316\\u0315\\u0300B"};
    int   srclen;
    UChar source[10];
    UCollationElements *iter;

    coll = ucol_openRules(rule, rulelen, UNORM_NFD, UCOL_TERTIARY, &status);
    ucol_setAttribute(coll, UCOL_NORMALIZATION_MODE, UCOL_ON, &status);
    if (U_FAILURE(status)){
        log_err("ERROR: in creation of collator using ucol_openRules()\n %s\n",
              myErrorName(status));
        return;
    }

    srclen = u_unescape(testdata[0], source, 10);
    iter = ucol_openElements(coll, source, srclen - 1, &status);
    backAndForth(iter);
    ucol_closeElements(iter);
    
    srclen = u_unescape(testdata[1], source, 10);
    iter = ucol_openElements(coll, source, srclen - 1, &status);
    backAndForth(iter);
    ucol_closeElements(iter);
    
    while (count < 12) {
        UCollationElements *iter;
        UChar               source[10];
        int                 srclen = 0;

        srclen = u_unescape(testdata[count], source, 10);
        iter = ucol_openElements(coll, source, srclen - 1, &status);

        if (U_FAILURE(status)){
            log_err("ERROR: in creation of collator element iterator\n %s\n",
                  myErrorName(status));
            return;
        }
        backAndForth(iter);
        ucol_closeElements(iter);

        iter = ucol_openElements(coll, source, -1, &status);

        if (U_FAILURE(status)){
            log_err("ERROR: in creation of collator element iterator\n %s\n",
                  myErrorName(status));
            return;
        }
        backAndForth(iter);
        ucol_closeElements(iter);
        count ++;
    }
    ucol_close(coll);
}

/**
 * Test for CollationElementIterator.previous()
 *
 * @bug 4108758 - Make sure it works with contracting characters
 *
 */
static void TestPrevious()
{
    UCollator *coll=NULL;
    UChar rule[50];
    UChar *source;
    UCollator *c1, *c2, *c3;
    UCollationElements *iter;
    UErrorCode status = U_ZERO_ERROR;

    test1=(UChar*)malloc(sizeof(UChar) * 50);
    test2=(UChar*)malloc(sizeof(UChar) * 50);
    u_uastrcpy(test1, "What subset of all possible test cases?");
    u_uastrcpy(test2, "has the highest probability of detecting");
    coll = ucol_open("en_US", &status);

    iter=ucol_openElements(coll, test1, u_strlen(test1), &status);
    log_verbose("English locale testing back and forth\n");
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        ucol_close(coll);
        return;
    }
    /* A basic test to see if it's working at all */
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(coll);

    /* Test with a contracting character sequence */
    u_uastrcpy(rule, "&a,A < b,B < c,C, d,D < z,Z < ch,cH,Ch,CH");
    c1 = ucol_openRules(rule, u_strlen(rule), UCOL_NO_NORMALIZATION, UCOL_DEFAULT_STRENGTH, &status);

    log_verbose("Contraction rule testing back and forth with no normalization\n");
    
    if (c1 == NULL || U_FAILURE(status))
    {
        log_err("Couldn't create a RuleBasedCollator with a contracting sequence\n %s\n",
            myErrorName(status));
        return;
    }
    source=(UChar*)malloc(sizeof(UChar) * 20);
    u_uastrcpy(source, "abchdcba");
    iter=ucol_openElements(c1, source, u_strlen(source), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        return;
    }
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(c1);
    
    /* Test with an expanding character sequence */
    u_uastrcpy(rule, "&a < b < c/abd < d");
    c2 = ucol_openRules(rule, u_strlen(rule), UCOL_NO_NORMALIZATION, UCOL_DEFAULT_STRENGTH,  &status);
    log_verbose("Expansion rule testing back and forth with no normalization\n");
    if (c2 == NULL || U_FAILURE(status))
    {
        log_err("Couldn't create a RuleBasedCollator with a contracting sequence.\n %s\n",
            myErrorName(status));
        return;
    }
    u_uastrcpy(source, "abcd");
    iter=ucol_openElements(c2, source, u_strlen(source), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        return;
    }
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(c2);
    /* Now try both */
    u_uastrcpy(rule, "&a < b < c/aba < d < z < ch");
    c3 = ucol_openRules(rule, u_strlen(rule), UCOL_DEFAULT_NORMALIZATION,  UCOL_DEFAULT_STRENGTH, &status);
    log_verbose("Expansion/contraction rule testing back and forth with no normalization\n");

    if (c3 == NULL || U_FAILURE(status))
    {
        log_err("Couldn't create a RuleBasedCollator with a contracting sequence.\n %s\n",
            myErrorName(status));
        return;
    }
    u_uastrcpy(source, "abcdbchdc");
    iter=ucol_openElements(c3, source, u_strlen(source), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        return;
    }
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(c3);
    source[0] = 0x0e41;
    source[1] = 0x0e02;
    source[2] = 0x0e41;
    source[3] = 0x0e02;
    source[4] = 0x0e27;
    source[5] = 0x61;
    source[6] = 0x62;
    source[7] = 0x63;
    source[8] = 0;

    coll = ucol_open("th_TH", &status);
    log_verbose("Thai locale testing back and forth with normalization\n");
    iter=ucol_openElements(coll, source, u_strlen(source), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        return;
    }
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(coll);

    /* prev test */
    source[0] = 0x0061;
    source[1] = 0x30CF;
    source[2] = 0x3099;
    source[3] = 0x30FC;
    source[4] = 0;

    coll = ucol_open("ja_JP", &status);
    log_verbose("Japanese locale testing back and forth with normalization\n");
    iter=ucol_openElements(coll, source, u_strlen(source), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
        return;
    }
    backAndForth(iter);
    ucol_closeElements(iter);
    ucol_close(coll);

    free(source);
    free(test1);
    free(test2);
}

/**
 * Test for getOffset() and setOffset()
 */
static void TestOffset()
{
    UErrorCode status= U_ZERO_ERROR;
    UCollator *en_us=NULL;
    UCollationElements *iter, *pristine;
    int32_t offset;
    int32_t *orders;
    int32_t orderLength=0;
    test1=(UChar*)malloc(sizeof(UChar) * 50);
    test2=(UChar*)malloc(sizeof(UChar) * 50);
    u_uastrcpy(test1, "What subset of all possible test cases?");
    u_uastrcpy(test2, "has the highest probability of detecting");
    en_us = ucol_open("en_US", &status);
    log_verbose("Testing getOffset and setOffset for CollationElements\n");
    iter=ucol_openElements(en_us, test1, u_strlen(test1), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
    ucol_close(en_us);
        return;
    }
    /* Run all the way through the iterator, then get the offset */

    orders = getOrders(iter, &orderLength);

    offset = ucol_getOffset(iter);

    if (offset != u_strlen(test1))
    {
        log_err("offset at end != length %d vs %d\n", offset,
            u_strlen(test1) );
    }

    /* Now set the offset back to the beginning and see if it works */
    pristine=ucol_openElements(en_us, test1, u_strlen(test1), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
            myErrorName(status));
    ucol_close(en_us);
        return;
    }
    status = U_ZERO_ERROR;

    ucol_setOffset(iter, 0, &status);
    if (U_FAILURE(status))
    {
        log_err("setOffset failed. %s\n",    myErrorName(status));
    }
    else
    {
        assertEqual(iter, pristine);
    }

    ucol_closeElements(pristine);
    ucol_closeElements(iter);
    free(orders);
    ucol_close(en_us);
    free(test1);
    free(test2);
}

/**
 * Test for setText()
 */
static void TestSetText()
{
    int32_t c,i;
    UErrorCode status = U_ZERO_ERROR;
    UCollator *en_us=NULL;
    UCollationElements *iter1, *iter2;
    test1=(UChar*)malloc(sizeof(UChar) * 50);
    test2=(UChar*)malloc(sizeof(UChar) * 50);
    u_uastrcpy(test1, "What subset of all possible test cases?");
    u_uastrcpy(test2, "has the highest probability of detecting");
    en_us = ucol_open("en_US", &status);
    log_verbose("testing setText for Collation elements\n");
    iter1=ucol_openElements(en_us, test1, u_strlen(test1), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator1 using ucol_openElements()\n %s\n",
            myErrorName(status));
    ucol_close(en_us);
        return;
    }
    iter2=ucol_openElements(en_us, test2, u_strlen(test2), &status);
    if(U_FAILURE(status)){
        log_err("ERROR: in creation of collation element iterator2 using ucol_openElements()\n %s\n",
            myErrorName(status));
    ucol_close(en_us);
        return;
    }

    /* Run through the second iterator just to exercise it */
    c = ucol_next(iter2, &status);
    i = 0;

    while ( ++i < 10 && (c != UCOL_NULLORDER))
    {
        if (U_FAILURE(status))
        {
            log_err("iter2->next() returned an error. %s\n", myErrorName(status));
            ucol_closeElements(iter2);
            ucol_closeElements(iter1);
    ucol_close(en_us);
            return;
        }

        c = ucol_next(iter2, &status);
    }

    /* Now set it to point to the same string as the first iterator */
    ucol_setText(iter2, test1, u_strlen(test1), &status);
    if (U_FAILURE(status))
    {
        log_err("call to iter2->setText(test1) failed. %s\n", myErrorName(status));
    }
    else
    {
        assertEqual(iter1, iter2);
    }

    ucol_closeElements(iter2);
    ucol_closeElements(iter1);
    ucol_close(en_us);
    free(test1);
    free(test2);
}



static void backAndForth(UCollationElements *iter)
{
    /* Run through the iterator forwards and stick it into an array */
    int32_t index, o;
    UErrorCode status = U_ZERO_ERROR;
    int32_t orderLength = 0;
    int32_t *orders;
    orders= getOrders(iter, &orderLength);


    /* Now go through it backwards and make sure we get the same values */
    index = orderLength;
    ucol_reset(iter);

    /* synwee : changed */
    while ((o = ucol_previous(iter, &status)) != UCOL_NULLORDER)
    {
      if (o != orders[-- index])
      {
        if (o == 0)
          index ++;
        else
        {
          while (index > 0 && orders[-- index] == 0)
          {
          }
          if (o != orders[index])
          {
            log_err("Mismatch at index : 0x%x\n", index);
            return;
          }

        }
      }
    }

    while (index != 0 && orders[index - 1] == 0) {
      index --;
    }

    if (index != 0)
    {
        log_err("Didn't get back to beginning - index is %d\n", index);

        ucol_reset(iter);
        log_err("\nnext: ");
        if ((o = ucol_next(iter, &status)) != UCOL_NULLORDER)
        {
            log_err("Error at %x\n", o);
        }
        log_err("\nprev: ");
        if ((o = ucol_previous(iter, &status)) != UCOL_NULLORDER)
        {
            log_err("Error at %x\n", o);
        }
        log_verbose("\n");
    }

    free(orders);
}

/** @bug 4108762
 * Test for getMaxExpansion()
 */
static void TestMaxExpansion()
{
    UErrorCode          status = U_ZERO_ERROR;
    UCollator          *coll   ;/*= ucol_open("en_US", &status);*/
    UChar               ch     = 0;
    UCollationElements *iter   ;/*= ucol_openElements(coll, &ch, 1, &status);*/

    UChar rule[256];
    u_uastrcpy(rule, "&a < ab < c/aba < d < z < ch");
    coll = ucol_openRules(rule, u_strlen(rule), UCOL_DEFAULT_NORMALIZATION,
        UCOL_DEFAULT_STRENGTH, &status);
    iter = ucol_openElements(coll, &ch, 1, &status);

    while (ch < 0xFFFF && U_SUCCESS(status)) {
        int      count = 1;
        uint32_t order;
        ch++;
        ucol_setText(iter, &ch, 1, &status);
        order = ucol_previous(iter, &status);

        /* thai management */
        if (order == 0)
            order = ucol_previous(iter, &status);

        while (U_SUCCESS(status) &&
            ucol_previous(iter, &status) != UCOL_NULLORDER) {
            count ++;
        }

        if (U_FAILURE(status) && ucol_getMaxExpansion(iter, order) < count) {
            log_err("Failure at codepoint %d, maximum expansion count < %d\n",
                ch, count);
        }
    }

    ucol_closeElements(iter);
    ucol_close(coll);
}

/**
 * Return an integer array containing all of the collation orders
 * returned by calls to next on the specified iterator
 */
static int32_t* getOrders(UCollationElements *iter, int32_t *orderLength)
{
    UErrorCode status;
    int32_t order;
    int32_t maxSize = 100;
    int32_t size = 0;
    int32_t *temp;
    int32_t *orders =(int32_t*)malloc(sizeof(int32_t) * maxSize);
    status= U_ZERO_ERROR;


    while ((order=ucol_next(iter, &status)) != UCOL_NULLORDER)
    {
        if (size == maxSize)
        {
            maxSize *= 2;
            temp = (int32_t*)malloc(sizeof(int32_t) * maxSize);

            memcpy(temp, orders, size * sizeof(int32_t));
            free(orders);
            orders = temp;

        }

        orders[size++] = order;
    }

    if (maxSize > size)
    {
        if (size == 0) {
            size = 1;
            temp = (int32_t*)malloc(sizeof(int32_t) * size);
            temp[0] = 0;
        }
        else {
            temp = (int32_t*)malloc(sizeof(int32_t) * size);
            memcpy(temp, orders, size * sizeof(int32_t));
        }

        free(orders);
        orders = temp;
    }

    *orderLength = size;
    return orders;
}


static void assertEqual(UCollationElements *i1, UCollationElements *i2)
{
    int32_t c1, c2;
    int32_t count = 0;
    UErrorCode status = U_ZERO_ERROR;

    do
    {
        c1 = ucol_next(i1, &status);
        c2 = ucol_next(i2, &status);

        if (c1 != c2)
        {
            log_err("Error in iteration %d assetEqual between\n  %d  and   %d, they are not equal\n", count, c1, c2);
            break;
        }

        count += 1;
    }
    while (c1 != UCOL_NULLORDER);
}

/**
 * Testing iterators with extremely small buffers
 */
static void TestSmallBuffer()
{
    UErrorCode          status = U_ZERO_ERROR;
    UCollator          *coll;
    UCollationElements *testiter,
                       *iter;
    int32_t             count = 0;
    int32_t            *testorders,
                       *orders;

    UChar teststr[500];
    UChar str[] = {0x300, 0x31A, 0};
    /*
    creating a long string of decomposable characters,
    since by default the writable buffer is of size 256
    */
    while (count < 500) {
        if ((count & 1) == 0) {
            teststr[count ++] = 0x300;
        }
        else {
            teststr[count ++] = 0x31A;
        }
    }

    coll = ucol_open("th_TH", &status);
    testiter = ucol_openElements(coll, teststr, 500, &status);
    iter = ucol_openElements(coll, str, 2, &status);

    orders     = getOrders(iter, &count);
    if (count != 2) {
        log_err("Error collation elements size is not 2 for \\u0300\\u031A\n");
    }

    /*
    this will rearrange the string data to 250 characters of 0x300 first then
    250 characters of 0x031A
    */
    testorders = getOrders(testiter, &count);

    if (count != 500) {
        log_err("Error decomposition does not give the right sized collation elements\n");
    }

    while (count != 0) {
        /* UCA collation element for 0x0F76 */
        if ((count > 250 && testorders[-- count] != orders[1]) ||
            (count <= 250 && testorders[-- count] != orders[0])) {
            log_err("Error decomposition does not give the right collation element at %d count\n", count);
            break;
        }
    }

    free(testorders);
    free(orders);
    ucol_closeElements(testiter);
    ucol_closeElements(iter);
    ucol_close(coll);
}

/**
* Sniplets of code from genuca
*/
static int32_t hex2num(char hex) {
    if(hex>='0' && hex <='9') {
        return hex-'0';
    } else if(hex>='a' && hex<='f') {
        return hex-'a'+10;
    } else if(hex>='A' && hex<='F') {
        return hex-'A'+10;
    } else {
        return 0;
    }
}

/** 
* Getting codepoints from a string 
* @param str character string contain codepoints seperated by space and ended 
*        by a semicolon
* @param codepoints array for storage, assuming size > 5
* @return position at the end of the codepoint section
*/
static char * getCodePoints(char *str, UChar *codepoints) {
    char *pStartCP = str;
    char *pEndCP   = str + 4;
    int   count    = 0;

    *codepoints = (UChar)((hex2num(*pStartCP) << 12) | 
                          (hex2num(*(pStartCP + 1)) << 8) | 
                          (hex2num(*(pStartCP + 2)) << 4) | 
                          (hex2num(*(pStartCP + 3))));
    codepoints ++;
    while (*pEndCP != ';') {
        pStartCP = pEndCP + 1;
        *codepoints = (UChar)((hex2num(*pStartCP) << 12) | 
                          (hex2num(*(pStartCP + 1)) << 8) | 
                          (hex2num(*(pStartCP + 2)) << 4) | 
                          (hex2num(*(pStartCP + 3))));
        codepoints ++;
        pEndCP = pStartCP + 4;
    }
    *codepoints = 0;
    return pEndCP + 1;
}

/**
* Sniplets of code from genuca
*/
int32_t readElement(char **from, char *to, char separator, UErrorCode *status) {
    if (U_SUCCESS(*status)) {
        char    buffer[1024];
        int32_t i = 0;
        while (**from != separator) {
            if (**from != ' ') {
                *(buffer+i++) = **from;
            }
            (*from)++;
        }
        (*from)++;
        *(buffer + i) = 0;
        strcpy(to, buffer);
        return i/2;            
    }
    
    return 0;
}

/**
* Sniplets of code from genuca
*/
uint32_t getSingleCEValue(char *primary, char *secondary, char *tertiary, 
                          UErrorCode *status) {
    if (U_SUCCESS(*status)) {
        uint32_t  value    = 0;
        char      primsave = '\0';
        char      secsave  = '\0';
        char      tersave  = '\0';
        char     *primend  = primary+4;
        char     *secend   = secondary+2;
        char     *terend   = tertiary+2;
        uint32_t  primvalue;
        uint32_t  secvalue;
        uint32_t  tervalue;

        if (strlen(primary) > 4) {
            primsave = *primend;
            *primend = '\0';
        }
        
        if (strlen(secondary) > 2) {
            secsave = *secend;
            *secend = '\0';
        }
        
        if (strlen(tertiary) > 2) {
            tersave = *terend;
            *terend = '\0';
        }

        primvalue = (*primary!='\0')?strtoul(primary, &primend, 16):0;
        secvalue  = (*secondary!='\0')?strtoul(secondary, &secend, 16):0;
        tervalue  = (*tertiary!='\0')?strtoul(tertiary, &terend, 16):0;
        if(primvalue <= 0xFF) {
          primvalue <<= 8;
        }

        value = ((primvalue << UCOL_PRIMARYORDERSHIFT) & UCOL_PRIMARYORDERMASK) 
           | ((secvalue << UCOL_SECONDARYORDERSHIFT) & UCOL_SECONDARYORDERMASK) 
           | (tervalue & UCOL_TERTIARYORDERMASK);

        if(primsave!='\0') {
            *primend = primsave;
        }
        if(secsave!='\0') {
            *secend = secsave;
        }
        if(tersave!='\0') {
            *terend = tersave;
        }
        return value;
    }
    return 0;
}

/** 
* Getting collation elements generated from a string 
* @param str character string contain collation elements contained in [] and 
*        seperated by space
* @param ce array for storage, assuming size > 20
* @param status error status
* @return position at the end of the codepoint section
*/
static char * getCEs(char *str, uint32_t *ces, UErrorCode *status) {
    char       *pStartCP     = uprv_strchr(str, '[');
    int         count        = 0;
    char       *pEndCP;
    char        primary[100];
    char        secondary[100];
    char        tertiary[100];
    
    while (*pStartCP == '[') {
        uint32_t primarycount   = 0;
        uint32_t secondarycount = 0;
        uint32_t tertiarycount  = 0;
        uint32_t CEi = 1;
        pEndCP = strchr(pStartCP, ']');
        if(pEndCP == NULL) {
            break;
        }
        pStartCP ++;

        primarycount   = readElement(&pStartCP, primary, ',', status);
        secondarycount = readElement(&pStartCP, secondary, ',', status);
        tertiarycount  = readElement(&pStartCP, tertiary, ']', status);

        /* I want to get the CEs entered right here, including continuation */
        ces[count ++] = getSingleCEValue(primary, secondary, tertiary, status);
        if (U_FAILURE(*status)) {
            break;
        }

        while (2 * CEi < primarycount || CEi < secondarycount || 
               CEi < tertiarycount) {
            uint32_t value = UCOL_CONTINUATION_MARKER; /* Continuation marker */
            if (2 * CEi < primarycount) {
                value |= ((hex2num(*(primary + 4 * CEi)) & 0xF) << 28);
                value |= ((hex2num(*(primary + 4 * CEi + 1)) & 0xF) << 24);
            }

            if (2 * CEi + 1 < primarycount) {
                value |= ((hex2num(*(primary + 4 * CEi + 2)) & 0xF) << 20);
                value |= ((hex2num(*(primary + 4 * CEi + 3)) &0xF) << 16);
            }

            if (CEi < secondarycount) {
                value |= ((hex2num(*(secondary + 2 * CEi)) & 0xF) << 12);
                value |= ((hex2num(*(secondary + 2 * CEi + 1)) & 0xF) << 8);
            }

            if (CEi < tertiarycount) {
                value |= ((hex2num(*(tertiary + 2 * CEi)) & 0x3) << 4);
                value |= (hex2num(*(tertiary + 2 * CEi + 1)) & 0xF);
            }

            CEi ++;
            ces[count ++] = value;
        }

      pStartCP = pEndCP + 1;
    }
    ces[count] = 0;
    return pStartCP;
}

/**
* Testing the CEs returned by the iterator
*/
static void TestCEs() {
    char        dir[150];        
    FileStream *file = NULL;
    char        line[300];
    char       *pDir = dir;
    char       *str;
    UChar       codepoints[5];
    uint32_t    ces[20];
    UErrorCode  status = U_ZERO_ERROR;
    UCollator          *coll = ucol_open("", &status);

    if (U_FAILURE(status)) {
        log_err("Error in opening root collator\n");
        return;
    }

    uprv_strcpy(pDir, getenv("ICU_DATA"));
    pDir += uprv_strlen(pDir);
    if (*(pDir - 1) != U_FILE_SEP_CHAR) {
        *pDir = U_FILE_SEP_CHAR;
        pDir ++;
    }

    /* dirty : because some platforms might not return the full path */
#ifdef XP_MAC
    uprv_strcpy(pDir, "..:..:data:unidata:FractionalUCA.txt");
#elif defined(WIN32) || defined(OS2)
    uprv_strcpy(pDir, "..\\..\\data\\unidata\\FractionalUCA.txt");
#else 
    uprv_strcpy(pDir, "../../data/unidata/FractionalUCA.txt");
#endif
    
    file = T_FileStream_open(dir, "r");

    if (file == NULL) {
        log_err("*** unable to open input FractionalUCA.txt file ***\n");
        return;
    }

    while (T_FileStream_readLine(file, line, sizeof(line)) != NULL) {
        int                 count = 0;
        UCollationElements *iter;
        /* skip this line if it is empty or a comment or is a return value 
        or start of some variable section */
        if(line[0] == 0 || line[0] == '#' || line[0] == '\n' || 
            line[0] == '[') {
            continue;
        }

        str = getCodePoints(line, codepoints);
        getCEs(str, ces, &status);
        if (U_FAILURE(status)) {
            log_err("Error in parsing collation elements in FractionalUCA.txt\n");
            break;
        }
        iter = ucol_openElements(coll, codepoints, -1, &status);      
        if (U_FAILURE(status)) {
            log_err("Error in opening collation elements\n");
            break;
        }     
        while (TRUE) {
            uint32_t ce = (uint32_t)ucol_next(iter, &status);
            if (ce == 0xFFFFFFFF) {
                ce = 0;
            }
            if (ce != ces[count] || U_FAILURE(status)) {
                log_err("Collation elements in FractionalUCA.txt and iterators do not match!\n");
                break;
            }
            if (ces[count] == 0) {
                break;
            }
            count ++;
        }
        ucol_closeElements(iter);
    }

    T_FileStream_close(file);
    ucol_close(coll);
}

/**
* Testing the discontigous contractions
*/
static void TestDiscontiguos() {
    const char               *rulestr    = 
                            "&z < AB < X\\u0300 < ABC < X\\u0300\\u0315";
          UChar               rule[50];
          int                 rulelen = u_unescape(rulestr, rule, 50);
    const char               *src[] = {
     "ADB", "ADBC", "A\\u0315B", "A\\u0315BC", 
    /* base character blocked */
     "XD\\u0300", "XD\\u0300\\u0315", 
    /* non blocking combining character */
     "X\\u0319\\u0300", "X\\u0319\\u0300\\u0315", 
     /* blocking combining character */
     "X\\u0314\\u0300", "X\\u0314\\u0300\\u0315", 
     /* contraction prefix */
     "ABDC", "AB\\u0315C","X\\u0300D\\u0315", "X\\u0300\\u0319\\u0315",
     "X\\u0300\\u031A\\u0315", 
     /* ends not with a contraction character */
     "X\\u0319\\u0300D", "X\\u0319\\u0300\\u0315D", "X\\u0300D\\u0315D", 
     "X\\u0300\\u0319\\u0315D", "X\\u0300\\u031A\\u0315D"
    };                              
    const char               *tgt[] = {
     /* non blocking combining character */
     "A D B", "A D BC", "A \\u0315 B", "A \\u0315 BC", 
    /* base character blocked */
     "X D \\u0300", "X D \\u0300\\u0315", 
    /* non blocking combining character */
     "X\\u0300 \\u0319", "X\\u0300\\u0315 \\u0319", 
     /* blocking combining character */
     "X \\u0314 \\u0300", "X \\u0314 \\u0300\\u0315", 
     /* contraction prefix */
     "AB DC", "AB \\u0315 C","X\\u0300 D \\u0315", "X\\u0300\\u0315 \\u0319",
     "X\\u0300 \\u031A \\u0315",
     /* ends not with a contraction character */
     "X\\u0300 \\u0319D", "X\\u0300\\u0315 \\u0319D", "X\\u0300 D\\u0315D", 
     "X\\u0300\\u0315 \\u0319D", "X\\u0300 \\u031A\\u0315D"
    };
          int                 size   = 20;
          UCollator          *coll;
          UErrorCode          status    = U_ZERO_ERROR;
          int                 count     = 0;
          UCollationElements *iter;      
          UCollationElements *resultiter;      
    
    coll       = ucol_openRules(rule, u_strlen(rule), UCOL_NO_NORMALIZATION,  
                          UCOL_DEFAULT_STRENGTH, &status);
    iter       = ucol_openElements(coll, rule, 1, &status);
    resultiter = ucol_openElements(coll, rule, 1, &status);

    if (U_FAILURE(status)) {
        log_err("Error opening collation rules\n");
        return;
    }

    while (count < size) {
        UChar  str[20];
        UChar  tstr[20];
        int    strlen = u_unescape(src[count], str, 20);
        UChar *s;

        ucol_setText(iter, str, strlen, &status);        
        if (U_FAILURE(status)) {
            log_err("Error opening collation iterator\n");
            return;
        }

        u_unescape(tgt[count], tstr, 20);
        s = tstr;

        log_verbose("count %d\n", count);
        
        while (TRUE) {
            uint32_t  ce;
            UChar    *e = u_strchr(s, ' ');
            if (e == 0) {
                e = u_strchr(s, 0);
            }
            ucol_setText(resultiter, s, e - s, &status);      
            ce = ucol_next(resultiter, &status);
            if (U_FAILURE(status)) {
                log_err("Error manipulating collation iterator\n");
                return;
            }
            while (ce != UCOL_NULLORDER) {
                if (ce != (uint32_t)ucol_next(iter, &status) || 
                    U_FAILURE(status)) {
                    log_err("Discontiguos contraction test mismatch\n");
                    return;
                }
                ce = ucol_next(resultiter, &status);
                if (U_FAILURE(status)) {
                    log_err("Error getting next collation element\n");
                    return;
                }
            }
            s = e + 1;
            if (*e == 0) {
                break;
            }
        }
        ucol_reset(iter);
        backAndForth(iter);
        count ++;
    }
    ucol_closeElements(resultiter);
    ucol_closeElements(iter);
    ucol_close(coll);
}